# Gemini 3.8 FlashHigh

> Source: Benchize (https://benchize.com) — https://benchize.com/models/gemini-3-8-flash-high. When citing or reusing this content, credit "Benchize" and link to https://benchize.com/models/gemini-3-8-flash-high.

Google

Canonical URL: https://benchize.com/models/gemini-3-8-flash-high
Last updated: Wed, 21 Oct 2026 18:05:10 GMT

Reproduction and quotation must name Benchize, link to the canonical HTML URL, and retain individual source citations.

Google

## Benchmarks

| Benchmark | Score | ± | Rank | Cost | Latency | Source |  |
| --- | --- | --- | --- | --- | --- | --- | --- |
| [Short Story Writing — Pairwise Comparison](https://benchize.com/benchmarks/lechmazur-short-story%3Apairwise.md) | 0.15 | 95% CI 0.00824–0.28 | 29 / 50 | — | — | [lechmazur short story · 8ba548d1cffefa5c472d3832e05f64e119a88fa4](https://api.github.com/repos/lechmazur/writing/contents/data/writing_gemini_argon_20261013/ratings.csv?ref=8ba548d1cffefa5c472d3832e05f64e119a88fa4) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `thurstone_rating` \| 0.1482517060222879 \| index \| pooled_observed_panel \| Higher is better \| \| `win_probability` \| 0.5221360708134836 \| score_fraction \| comparison_pool \| Higher is better \| ### Configuration benchmarkScope data/writing_gemini_argon_20261013 Reasoning effort high thinkingEnabled true ### Reported statistics Uncertainty 95% confidence interval: 0.00824–0.28 bootstrapSamples 300 confidenceInterval95 ["0.008239475449616123","0.27563638384343925"] sourceRank 29 ### Source [Original source](https://api.github.com/repos/lechmazur/writing/contents/data/writing_gemini_argon_20261013/ratings.csv?ref=8ba548d1cffefa5c472d3832e05f64e119a88fa4) Source lechmazur short story Original model label Gemini 3.8 Flash (high) Source version 8ba548d1cffefa5c472d3832e05f64e119a88fa4 Observed 2026-10-21T16:34:21.130Z Evidence SHA-256 `c36baebed9f302ff7ad8fe97907d72e1757a015a39b6a3b23437597cf8e3bff6` Source locator `/rows/28` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / international trade sanctions](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Ainternational_trade_sanctions.md) | 0% | SE ±0 pp | 4 / 49 | $3.92 | 859.2source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 0 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 3.921448 \| USD \| per_test \| Lower is better \| \| `latency` \| 859.204 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 81 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/international_trade_sanctions/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Arena text style control / math](https://benchize.com/benchmarks/lmarena%3Atext_style_control%3Amath.md) | 1,526.3 | — | 3 / 149 | — | — | [Arena](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text_style_control/latest) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `rating` \| 1526.2953339165267 \| rating \| reported \| Higher is better \| ### Configuration category math config text_style_control Reasoning effort high thinkingEnabled true ### Reported statistics lower 1512.8659871369218 rank 3 upper 1539.7246806961318 variance 46.947621553391436 votes 2094 ### Source [Original source](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text_style_control/latest) Source Arena Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:40:19.533Z Evidence SHA-256 `d4875c03cb7ec0c6e8943c3145fd261225e8841b937240dc62a3032c4a798d84` Source locator `/rows/8915/row` |  |  |  |  |  |  |  |
| [Arena webdev / image to webdev](https://benchize.com/benchmarks/lmarena%3Awebdev%3Aimage_to_webdev.md) | 1,592.72 | — | 17 / 42 | — | — | [Arena](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/webdev/latest) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `rating` \| 1592.7182186745918 \| rating \| reported \| Higher is better \| ### Configuration category image_to_webdev config webdev Reasoning effort high thinkingEnabled true ### Reported statistics lower 1580.1632074366007 rank 19 upper 1605.2732299125828 variance 41.03344967201824 votes 2516 ### Source [Original source](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/webdev/latest) Source Arena Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:40:19.533Z Evidence SHA-256 `f82507859808110820a4ab6df15561b8864137ddefc1f6fb0c552af2e58ea738` Source locator `/rows/166/row` |  |  |  |  |  |  |  |
| [Terminal-Bench 4.0 / hardware](https://benchize.com/benchmarks/vals-ai%3Aterminal-bench-4%3Ahardware.md) | 26.67% | SE ±6.67 pp | 11 / 36 | — | — | [Vals AI · 4.0](https://www.vals.ai/_astro/benchmark_view_terminal_bench_4.CAnGyzC_.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 26.667 \| percent \| reported \| Higher is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±6.67 pp sourcePopulation 52 stderr 6.667 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_terminal_bench_4.CAnGyzC_.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 4.0 Observed 2026-10-21T16:46:10.642Z Evidence SHA-256 `98f6964580e901f1aa143db664bd92aa558415f643569492545d67c47d7f519b` Source locator `/tasks/hardware/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [CUA-bench / efootball](https://benchize.com/benchmarks/vals-ai%3Acua_bench%3Aefootball.md) | 0% | SE ±0 pp | 3 / 8 | $67.98 | 34.4source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_cua_bench.DpgrKYWA.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 0 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 67.980694 \| USD \| per_test \| Lower is better \| \| `latency` \| 34.4 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort high harness google-computer-use maxOutputTokens 128000 provider google reasoning true Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 9 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_cua_bench.DpgrKYWA.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:46:16.228Z Evidence SHA-256 `8c415c63f28cd80496a4ccca27d1062ec98bb55268e58cba105b8cf2d7ee53d3` Source locator `/tasks/efootball/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [SRE Bench / partial](https://benchize.com/benchmarks/vals-ai%3Asrebench%3Apartial.md) | 19.66% | SE ±2.09 pp | 11 / 24 | $27.46 | 8,853.86source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_sre_bench.Bd0uBJxH.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 19.656 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 27.459572 \| USD \| per_test \| Lower is better \| \| `latency` \| 8853.856 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.09 pp sourcePopulation 35 stderr 2.087 tokenTotals {"cache_read_tokens":36219215322,"cache_write_tokens":null,"input_tokens":37086688061,"output_tokens":61375519,"reasoning_tokens":34869435} ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_sre_bench.Bd0uBJxH.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:46:15.724Z Evidence SHA-256 `a10b2c964f5878b0fc21c888703f311a5ef7d7a74fc3c529cab3ddd9fa474e2b` Source locator `/tasks/partial/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [SWE-bench Multimodal / prettier](https://benchize.com/benchmarks/vals-ai%3Aswebench-multimodal%3Aprettier.md) | 46.15% | SE ±14.39 pp | 12 / 14 | — | — | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_swebench_multimodal.C7UOLrm7.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 46.154 \| percent \| reported \| Higher is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±14.39 pp sourcePopulation 20 stderr 14.391 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_swebench_multimodal.C7UOLrm7.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:46:12.211Z Evidence SHA-256 `5302c42065b017d7deaa43d9e34be1ec69893b7ec4cceb6c07bfbabbd20b1f82` Source locator `/tasks/prettier/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [LegalBench / rule tasks](https://benchize.com/benchmarks/vals-ai%3Alegal_bench%3Arule_tasks.md) | 87.08% | SE ±1.24 pp | 6 / 44 | $0.00785 | 3.74source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 87.076 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.007846 \| USD \| per_test \| Lower is better \| \| `latency` \| 3.736 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.24 pp sourcePopulation 151 stderr 1.242 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/rule_tasks/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Arena webdev / webdev](https://benchize.com/benchmarks/lmarena%3Awebdev%3Awebdev.md) | 1,584.7 | — | 25 / 90 | — | — | [Arena](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/webdev/latest) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `rating` \| 1584.6956846722271 \| rating \| reported \| Higher is better \| ### Configuration category webdev config webdev Reasoning effort high thinkingEnabled true ### Reported statistics lower 1577.3388914417835 rank 35 upper 1592.0524779026705 variance 14.089024290443556 votes 10851 ### Source [Original source](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/webdev/latest) Source Arena Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:40:19.533Z Evidence SHA-256 `f82507859808110820a4ab6df15561b8864137ddefc1f6fb0c552af2e58ea738` Source locator `/rows/246/row` |  |  |  |  |  |  |  |
| [Tax Agent Bench / weighted pass numbers calculations](https://benchize.com/benchmarks/vals-ai%3Atax_agent_bench%3Aweighted_pass__numbers_calculations.md) | 83.05% | SE ±6.13 pp | 8 / 46 | $0.62 | 120.6source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_tax_agent_bench.BFpexnpS.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 83.051 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.619008 \| USD \| per_test \| Lower is better \| \| `latency` \| 120.595 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±6.13 pp sourcePopulation 73 stderr 6.134 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_tax_agent_bench.BFpexnpS.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:55.958Z Evidence SHA-256 `446db182b67efb7923b3187a4040ec3a3ea492b8ad699a2b6d708d714cc063ac` Source locator `/tasks/weighted_pass__numbers_calculations/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Arena text style control / overall](https://benchize.com/benchmarks/lmarena%3Atext_style_control%3Aoverall.md) | 1,496.31 | — | 9 / 151 | — | — | [Arena](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text_style_control/latest) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `rating` \| 1496.30827195144 \| rating \| reported \| Higher is better \| ### Configuration category overall config text_style_control Reasoning effort high thinkingEnabled true ### Reported statistics lower 1491.7813303212706 rank 9 upper 1500.8352135816092 variance 5.3347442936423874 votes 41423 ### Source [Original source](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text_style_control/latest) Source Arena Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:40:19.533Z Evidence SHA-256 `d4875c03cb7ec0c6e8943c3145fd261225e8841b937240dc62a3032c4a798d84` Source locator `/rows/8/row` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / overall](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Aoverall.md) | 10% | SE ±2.42 pp | 10 / 49 | $3.66 | 1,760.87source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 10 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 3.656207 \| USD \| per_test \| Lower is better \| \| `latency` \| 1760.871 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.42 pp sourcePopulation 81 stderr 2.415 tokenTotals {"cache_read_tokens":1066501383,"cache_write_tokens":null,"input_tokens":1171742866,"output_tokens":16120993,"reasoning_tokens":5514918} ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/overall/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Mystery Game Puzzles](https://benchize.com/benchmarks/epoch-ai%3Amystery_game_puzzles.md) | 47% | SE ±5.02 pp | 13 / 71 | — | — | [epoch ai · Mystery Game Puzzles](https://epoch.ai/data/benchmark_data.zip#file=mystery_game_puzzles.csv) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 0.47 \| score_fraction \| Best score (across scorers) \| Higher is better \| \| `mean_score` \| 0.47 \| score_fraction \| mean_score \| Higher is better \| ### Configuration benchmarkVersion Mystery Game Puzzles reasoningBudget Not reported Reasoning effort high sourceModelVersion gemini-3.8-flash_high ### Reported statistics Uncertainty Standard error: ±5.02 pp standardError 0.05016135580465918 ### Source [Original source](https://epoch.ai/data/benchmark_data.zip#file=mystery_game_puzzles.csv) Source epoch ai Original model label gemini-3.8-flash Source version Mystery Game Puzzles Evaluated 2026-09-02T17:15:36.000Z Observed 2026-10-21T16:30:24.199Z Evidence SHA-256 `795b977bc82c460a5e2c1d6dd7b2569ddb21853f3829c4f6fdd16edfaf47bfcf` Source locator `/rows/13` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / healthcare life sciences](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Ahealthcare_life_sciences.md) | 20% | SE ±20 pp | 5 / 49 | $3.47 | 833.11source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 20 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 3.474289 \| USD \| per_test \| Lower is better \| \| `latency` \| 833.11 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±20 pp sourcePopulation 81 stderr 20 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/healthcare_life_sciences/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [GPQA Diamond](https://benchize.com/benchmarks/gpqa-diamond.md) | 95.39% | SE ±1.4 pp | 4 / 125 | — | — | [epoch ai · GPQA diamond](https://epoch.ai/data/benchmark_data.zip#file=gpqa_diamond.csv) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 0.9539141414141414 \| score_fraction \| Best score (across scorers) \| Higher is better \| \| `mean_score` \| 0.9539141414141414 \| score_fraction \| mean_score \| Higher is better \| ### Configuration benchmarkVersion GPQA diamond reasoningBudget Not reported Reasoning effort high sourceModelVersion gemini-3.8-flash_high ### Reported statistics Uncertainty Standard error: ±1.4 pp standardError 0.01401136318210652 ### Source [Original source](https://epoch.ai/data/benchmark_data.zip#file=gpqa_diamond.csv) Source epoch ai Original model label gemini-3.8-flash Source version GPQA diamond Evaluated 2026-09-02T17:15:36.000Z Observed 2026-10-21T16:30:24.199Z Evidence SHA-256 `4b2dd8376cdd249a2c3065603575208fa726c15eabbb1622689a9abc146ba29a` Source locator `/rows/12` |  |  |  |  |  |  |  |
| [Arena webdev / overall](https://benchize.com/benchmarks/lmarena%3Awebdev%3Aoverall.md) | 1,584.7 | — | 25 / 90 | — | — | [Arena](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/webdev/latest) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `rating` \| 1584.6956846722271 \| rating \| reported \| Higher is better \| ### Configuration category overall config webdev Reasoning effort high thinkingEnabled true ### Reported statistics lower 1577.3388914417835 rank 35 upper 1592.0524779026705 variance 14.089024290443556 votes 10851 ### Source [Original source](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/webdev/latest) Source Arena Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:40:19.533Z Evidence SHA-256 `f82507859808110820a4ab6df15561b8864137ddefc1f6fb0c552af2e58ea738` Source locator `/rows/34/row` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / criteria pass rate data privacy cybersecurity](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Acriteria_pass_rate__data_privacy_cybersecurity.md) | 93.29% | SE ±1.49 pp | 17 / 49 | $2.44 | 682.68source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 93.286 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 2.439638 \| USD \| per_test \| Lower is better \| \| `latency` \| 682.68 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.49 pp sourcePopulation 81 stderr 1.49 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__data_privacy_cybersecurity/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [AutomationBench](https://benchize.com/benchmarks/zapier%3Aautomationbench.md) | 29.68% | — | 18 / 113 | $0.62 | — | [zapier](https://framerusercontent.com/sites/4WTSl4BNjd1q9QFEFibC6h/lOvLnhItj8S8-1vyjH_s33Jn-WmUXPmiJ5Diajl0Q38.CMutaHYl.mjs) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 29.68 \| percent \| reported \| Higher is better \| \| `cost` \| 0.62 \| USD \| per_task \| Lower is better \| ### Configuration benchmarkVersion 1.0.6 costFootnote true Fallback policy Not reported Reasoning effort high Model configuration Gemini 3.8 Flash (High) thinkingEnabled true ### Reported statistics rank 19 ### Source [Original source](https://framerusercontent.com/sites/4WTSl4BNjd1q9QFEFibC6h/lOvLnhItj8S8-1vyjH_s33Jn-WmUXPmiJ5Diajl0Q38.CMutaHYl.mjs) Source zapier Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:34:56.783Z Evidence SHA-256 `43978e38f018789fa59c11de5328abe9e5905bfa9a5a4d601fa3aa4d9c153b0d` Source locator `/rows/18` |  |  |  |  |  |  |  |
| [τ³-Banking](https://benchize.com/benchmarks/mercor%3Atau3-banking.md) | 48.8% | — | 6 / 24 | — | — | [mercor](https://www.mercor.com/apex/partner-benchmarks/tau-3-banking-leaderboard/) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `pass_1` \| 48.8 \| percent \| task_completion_pass_at_1 \| Higher is better \| ### Configuration datasetUrl Not reported datasetVersion Not reported harness Not reported Reasoning effort high ### Reported statistics reportedErrors {"pass_1":8.8} sampleSize 291 ### Source [Original source](https://www.mercor.com/apex/partner-benchmarks/tau-3-banking-leaderboard/) Source mercor Original model label Gemini 3.8 Flash Observed 2026-10-21T16:33:27.812Z Evidence SHA-256 `78850adeb8ad047fcb1dd65cc04ffaca61c86fc67e7035f9973805b02483ed07` Source locator `/__NEXT_DATA__/props/pageProps/leaderboard/rows/2` |  |  |  |  |  |  |  |
| [SRE Bench / solved](https://benchize.com/benchmarks/vals-ai%3Asrebench%3Asolved.md) | 11.07% | SE ±1.94 pp | 10 / 24 | $27.46 | 8,853.86source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_sre_bench.Bd0uBJxH.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 11.069 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 27.459572 \| USD \| per_test \| Lower is better \| \| `latency` \| 8853.856 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.94 pp sourcePopulation 35 stderr 1.942 tokenTotals {"cache_read_tokens":36219215322,"cache_write_tokens":null,"input_tokens":37086688061,"output_tokens":61375519,"reasoning_tokens":34869435} ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_sre_bench.Bd0uBJxH.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:46:15.724Z Evidence SHA-256 `a10b2c964f5878b0fc21c888703f311a5ef7d7a74fc3c529cab3ddd9fa474e2b` Source locator `/tasks/solved/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [EMB v1.1 / dataroom](https://benchize.com/benchmarks/vals-ai%3Aemb-v1-1%3Adataroom.md) | 87.8% | SE ±3.46 pp | 2 / 19 | $8.51 | 955.74source duration | [Vals AI · 1.1](https://www.vals.ai/_astro/benchmark_view_emb_v1_1.11L06J7L.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 87.797 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 8.512634 \| USD \| per_test \| Lower is better \| \| `latency` \| 955.744 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±3.46 pp sourcePopulation 30 stderr 3.459 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_emb_v1_1.11L06J7L.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1.1 Observed 2026-10-21T16:45:55.456Z Evidence SHA-256 `eea2ef2c83a6dc7d553c7c6531c287710ea7ffbc56ee1ccec06c85b104b2408e` Source locator `/tasks/dataroom/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Arena text / industry writing and literature and language](https://benchize.com/benchmarks/lmarena%3Atext%3Aindustry_writing_and_literature_and_language.md) | 1,493.78 | — | 8 / 151 | — | — | [Arena](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text/latest) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `rating` \| 1493.7755979235694 \| rating \| reported \| Higher is better \| ### Configuration category industry_writing_and_literature_and_language config text Reasoning effort high thinkingEnabled true ### Reported statistics lower 1486.8817928310414 rank 8 upper 1500.6694030160977 variance 12.371484603128007 votes 11700 ### Source [Original source](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text/latest) Source Arena Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:40:19.533Z Evidence SHA-256 `8e027ef69b8de06c4b2a8b217bfe816b1f1536269bb257fd295b4ef1ba8246ac` Source locator `/rows/7125/row` |  |  |  |  |  |  |  |
| [MMLU Pro / engineering](https://benchize.com/benchmarks/vals-ai%3Ammlu_pro%3Aengineering.md) | 88.75% | SE ±1.02 pp | 5 / 40 | $0.05 | 14.79source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 88.751 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.048753 \| USD \| per_test \| Lower is better \| \| `latency` \| 14.793 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.02 pp sourcePopulation 138 stderr 1.019 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:46:06.503Z Evidence SHA-256 `dab025507f46c716dd8d145e90ec0a737d9c56219853f41e4b1cc3c5e8c222b4` Source locator `/tasks/engineering/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [HLAB-Vals-1.1 / structured finance securitization](https://benchize.com/benchmarks/vals-ai%3Ahlab%3Astructured_finance_securitization.md) | 0% | SE ±0 pp | 9 / 14 | $3.3 | 1,084.01source duration | [Vals AI · 1.1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 0 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 3.302349 \| USD \| per_test \| Lower is better \| \| `latency` \| 1084.012 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 24 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1.1 Observed 2026-10-21T16:45:52.890Z Evidence SHA-256 `14681f4a2310e0f257df9ec07c9c81e5dadc8f6f71634f02e174a02a8096d868` Source locator `/tasks/structured_finance_securitization/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / criteria pass rate litigation dispute resolution](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Acriteria_pass_rate__litigation_dispute_resolution.md) | 89.89% | SE ±1.85 pp | 8 / 49 | $4.59 | 2,372.29source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 89.888 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 4.594796 \| USD \| per_test \| Lower is better \| \| `latency` \| 2372.292 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.85 pp sourcePopulation 81 stderr 1.849 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__litigation_dispute_resolution/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [LiveCodeBench / hard](https://benchize.com/benchmarks/vals-ai%3Alcb%3Ahard.md) | 78% | SE ±2.21 pp | 3 / 43 | $0.15 | 49.71source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_lcb.miyRUvVa.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 78 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.14965 \| USD \| per_test \| Lower is better \| \| `latency` \| 49.71 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.21 pp sourcePopulation 143 stderr 2.214 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_lcb.miyRUvVa.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:46:14.220Z Evidence SHA-256 `1a7c7dabc9d95a0182886b1edc42623987cac2095a0d407a5e334f37acdfa2f5` Source locator `/tasks/hard/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / criteria pass rate arbitration international dispute resolution](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Acriteria_pass_rate__arbitration_international_dispute_resolution.md) | 89.61% | SE ±1.83 pp | 13 / 49 | $4.27 | 2,578.72source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 89.606 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 4.266851 \| USD \| per_test \| Lower is better \| \| `latency` \| 2578.722 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.83 pp sourcePopulation 81 stderr 1.83 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__arbitration_international_dispute_resolution/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Arena vision style control / chinese](https://benchize.com/benchmarks/lmarena%3Avision_style_control%3Achinese.md) | 1,383.73 | — | 1 / 74 | — | — | [Arena](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/vision_style_control/latest) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `rating` \| 1383.7344522524572 \| rating \| reported \| Higher is better \| ### Configuration category chinese config vision_style_control Reasoning effort high thinkingEnabled true ### Reported statistics lower 1343.9762835254955 rank 1 upper 1423.4926209794187 variance 411.48742035347965 votes 224 ### Source [Original source](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/vision_style_control/latest) Source Arena Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:40:19.533Z Evidence SHA-256 `38568936995f36625aae49e9ea05557d55c1b1df9b2954f5c50cee523f26fe92` Source locator `/rows/201/row` |  |  |  |  |  |  |  |
| [Finance Agent v2.1 / general quantitative analysis](https://benchize.com/benchmarks/vals-ai%3Afabv2-1%3Ageneral_quantitative_analysis.md) | 80.18% | SE ±4.98 pp | 7 / 18 | $1.99 | 401.73source duration | [Vals AI · 2.1](https://www.vals.ai/_astro/benchmark_view_finance_agent_v2_1.ZbMhy9xo.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 80.178 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 1.985641 \| USD \| per_test \| Lower is better \| \| `latency` \| 401.728 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±4.98 pp sourcePopulation 28 stderr 4.977 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_finance_agent_v2_1.ZbMhy9xo.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 2.1 Observed 2026-10-21T16:45:54.954Z Evidence SHA-256 `f1b0c2d7dca4e73b52136862a9be059f7c8481c69f15d705ce90bf3bd9f78429` Source locator `/tasks/general_quantitative_analysis/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [ToneBench Scriptwriting](https://benchize.com/benchmarks/tonebench%3Ascripts.md) | 1,152.5 | Range 1,111.1–1,195.8 | 93 / 126 | — | — | [tonebench](https://benchmark.towardsai.com/data/leaderboard.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1152.5 \| rating \| reported \| Higher is better \| \| `overall` \| 70.22 \| percent \| reported \| Higher is better \| \| `anti_slop_numbers` \| 60.78 \| source_score \| rubric \| Higher is better \| \| `continuity_emotion` \| 73.24 \| source_score \| rubric \| Higher is better \| \| `cue_quality` \| 73.86 \| source_score \| rubric \| Higher is better \| \| `hook_strength` \| 74.29 \| source_score \| rubric \| Higher is better \| \| `length_adherence` \| 67.76 \| source_score \| rubric \| Higher is better \| \| `slop` \| 75.52 \| source_score \| rubric \| Higher is better \| \| `slop_score` \| 90.26 \| source_score \| rubric \| Higher is better \| \| `substance_accuracy` \| 73.57 \| source_score \| rubric \| Higher is better \| \| `tone_voice` \| 61.92 \| source_score \| rubric \| Higher is better \| \| `writing_quality` \| 68.17 \| source_score \| rubric \| Higher is better \| \| `youtube_best_practices` \| 72.75 \| source_score \| rubric \| Higher is better \| ### Configuration provider google Reasoning effort high scoreScope 10_task_current sourceConfiguration gemini-3-8-flash-high ### Reported statistics Uncertainty Reported interval: 1,111.1–1,195.8 interval [1111.1,1195.8] rank 96 runCount 5 ### Source [Original source](https://benchmark.towardsai.com/data/leaderboard.json) Source tonebench Original model label Gemini 3.8 Flash (high thinking) Evaluated 2026-10-11 Observed 2026-10-21T16:32:48.862Z Evidence SHA-256 `4264338337f418d3ded7ace250cad64fe12cacbd4b79a312bfaf0ae9268a01c3` Source locator `/rows/95` |  |  |  |  |  |  |  |
| [Arena text style control / expert](https://benchize.com/benchmarks/lmarena%3Atext_style_control%3Aexpert.md) | 1,526.11 | — | 15 / 148 | — | — | [Arena](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text_style_control/latest) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `rating` \| 1526.1096121122587 \| rating \| reported \| Higher is better \| ### Configuration category expert config text_style_control Reasoning effort high thinkingEnabled true ### Reported statistics lower 1516.85719420624 rank 16 upper 1535.3620300182774 variance 22.285085198999493 votes 4889 ### Source [Original source](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text_style_control/latest) Source Arena Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:40:19.533Z Evidence SHA-256 `d4875c03cb7ec0c6e8943c3145fd261225e8841b937240dc62a3032c4a798d84` Source locator `/rows/2486/row` |  |  |  |  |  |  |  |
| [Terminal-Bench 4.0 / media](https://benchize.com/benchmarks/vals-ai%3Aterminal-bench-4%3Amedia.md) | 16.67% | SE ±8.33 pp | 17 / 36 | — | — | [Vals AI · 4.0](https://www.vals.ai/_astro/benchmark_view_terminal_bench_4.CAnGyzC_.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 16.667 \| percent \| reported \| Higher is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±8.33 pp sourcePopulation 52 stderr 8.333 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_terminal_bench_4.CAnGyzC_.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 4.0 Observed 2026-10-21T16:46:10.642Z Evidence SHA-256 `98f6964580e901f1aa143db664bd92aa558415f643569492545d67c47d7f519b` Source locator `/tasks/media/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [HLAB-Vals-1.1 / criteria pass rate energy natural resources](https://benchize.com/benchmarks/vals-ai%3Ahlab%3Acriteria_pass_rate__energy_natural_resources.md) | 97.53% | SE ±0.89 pp | 5 / 14 | $4.41 | 1,422.04source duration | [Vals AI · 1.1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 97.525 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 4.405884 \| USD \| per_test \| Lower is better \| \| `latency` \| 1422.04 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.89 pp sourcePopulation 24 stderr 0.894 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1.1 Observed 2026-10-21T16:45:52.890Z Evidence SHA-256 `14681f4a2310e0f257df9ec07c9c81e5dadc8f6f71634f02e174a02a8096d868` Source locator `/tasks/criteria_pass_rate__energy_natural_resources/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / criteria pass rate corporate ma](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Acriteria_pass_rate__corporate_ma.md) | 91.35% | SE ±1.59 pp | 22 / 49 | $3.55 | 799.5source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 91.346 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 3.550379 \| USD \| per_test \| Lower is better \| \| `latency` \| 799.498 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.59 pp sourcePopulation 81 stderr 1.594 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__corporate_ma/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [EMB v1.1 / operating](https://benchize.com/benchmarks/vals-ai%3Aemb-v1-1%3Aoperating.md) | 69.19% | SE ±5.56 pp | 11 / 19 | $6.43 | 621.68source duration | [Vals AI · 1.1](https://www.vals.ai/_astro/benchmark_view_emb_v1_1.11L06J7L.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 69.19 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 6.429094 \| USD \| per_test \| Lower is better \| \| `latency` \| 621.677 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±5.56 pp sourcePopulation 30 stderr 5.563 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_emb_v1_1.11L06J7L.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1.1 Observed 2026-10-21T16:45:55.456Z Evidence SHA-256 `eea2ef2c83a6dc7d553c7c6531c287710ea7ffbc56ee1ccec06c85b104b2408e` Source locator `/tasks/operating/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [HLAB-Vals-1.1 / criteria pass rate environmental esg](https://benchize.com/benchmarks/vals-ai%3Ahlab%3Acriteria_pass_rate__environmental_esg.md) | 92.47% | SE ±1.58 pp | 4 / 14 | $3.22 | 1,210.54source duration | [Vals AI · 1.1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 92.473 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 3.219091 \| USD \| per_test \| Lower is better \| \| `latency` \| 1210.54 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.58 pp sourcePopulation 24 stderr 1.582 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1.1 Observed 2026-10-21T16:45:52.890Z Evidence SHA-256 `14681f4a2310e0f257df9ec07c9c81e5dadc8f6f71634f02e174a02a8096d868` Source locator `/tasks/criteria_pass_rate__environmental_esg/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [HLAB-Vals-1.1 / criteria pass rate immigration](https://benchize.com/benchmarks/vals-ai%3Ahlab%3Acriteria_pass_rate__immigration.md) | 92% | SE ±1.64 pp | 5 / 14 | $2.61 | 1,107.64source duration | [Vals AI · 1.1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 92 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 2.608087 \| USD \| per_test \| Lower is better \| \| `latency` \| 1107.642 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.64 pp sourcePopulation 24 stderr 1.639 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1.1 Observed 2026-10-21T16:45:52.890Z Evidence SHA-256 `14681f4a2310e0f257df9ec07c9c81e5dadc8f6f71634f02e174a02a8096d868` Source locator `/tasks/criteria_pass_rate__immigration/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [HLAB-Vals-1.1 / funds asset management](https://benchize.com/benchmarks/vals-ai%3Ahlab%3Afunds_asset_management.md) | 0% | SE ±0 pp | 10 / 14 | $3.57 | 1,365.91source duration | [Vals AI · 1.1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 0 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 3.568526 \| USD \| per_test \| Lower is better \| \| `latency` \| 1365.906 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 24 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1.1 Observed 2026-10-21T16:45:52.890Z Evidence SHA-256 `14681f4a2310e0f257df9ec07c9c81e5dadc8f6f71634f02e174a02a8096d868` Source locator `/tasks/funds_asset_management/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Arena text style control / industry entertainment and sports and media](https://benchize.com/benchmarks/lmarena%3Atext_style_control%3Aindustry_entertainment_and_sports_and_media.md) | 1,476.8 | — | 9 / 151 | — | — | [Arena](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text_style_control/latest) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `rating` \| 1476.7995582155445 \| rating \| reported \| Higher is better \| ### Configuration category industry_entertainment_and_sports_and_media config text_style_control Reasoning effort high thinkingEnabled true ### Reported statistics lower 1469.6824400666542 rank 9 upper 1483.9166763644348 variance 13.185972597804094 votes 10947 ### Source [Original source](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text_style_control/latest) Source Arena Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:40:19.533Z Evidence SHA-256 `d4875c03cb7ec0c6e8943c3145fd261225e8841b937240dc62a3032c4a798d84` Source locator `/rows/4701/row` |  |  |  |  |  |  |  |
| [Terminal-Bench-Science / engineering sciences](https://benchize.com/benchmarks/vals-ai%3Aterminal-bench-science%3Aengineering_sciences.md) | 11.11% | SE ±11.11 pp | 7 / 30 | — | — | [Vals AI · 0.1](https://www.vals.ai/_astro/benchmark_view_terminal_bench_science.o-HR8Nyt.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 11.111 \| percent \| reported \| Higher is better \| ### Configuration computeEffort Not reported harness Mini-SWE-agent maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±11.11 pp sourcePopulation 41 stderr 11.111 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_terminal_bench_science.o-HR8Nyt.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 0.1 Observed 2026-10-21T16:46:03.529Z Evidence SHA-256 `a30a8618deaecf018ea1c8c9f675576cdd9ab79742d05788067db49afcc5f91b` Source locator `/tasks/engineering_sciences/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Terminal-Bench-Science / mathematical sciences](https://benchize.com/benchmarks/vals-ai%3Aterminal-bench-science%3Amathematical_sciences.md) | 11.77% | SE ±8.06 pp | 10 / 30 | — | — | [Vals AI · 0.1](https://www.vals.ai/_astro/benchmark_view_terminal_bench_science.o-HR8Nyt.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 11.765 \| percent \| reported \| Higher is better \| ### Configuration computeEffort Not reported harness Mini-SWE-agent maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±8.06 pp sourcePopulation 41 stderr 8.055 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_terminal_bench_science.o-HR8Nyt.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 0.1 Observed 2026-10-21T16:46:03.529Z Evidence SHA-256 `a30a8618deaecf018ea1c8c9f675576cdd9ab79742d05788067db49afcc5f91b` Source locator `/tasks/mathematical_sciences/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Code Migration v1.1 / cli](https://benchize.com/benchmarks/vals-ai%3Acode-migration-v1-1%3Acli.md) | 27.62% | SE ±2.65 pp | 14 / 17 | $19.64 | 11,148.61source duration | [Vals AI · 1.1](https://www.vals.ai/_astro/benchmark_view_code_migration_v1_1.bJHk6HEY.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 27.62 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 19.642889 \| USD \| per_test \| Lower is better \| \| `latency` \| 11148.606 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.65 pp sourcePopulation 26 stderr 2.652 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_code_migration_v1_1.bJHk6HEY.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1.1 Observed 2026-10-21T16:46:09.631Z Evidence SHA-256 `352faac603c73f05cc01e89214fb4adcafa05ea7432ea71ea9903a6b5fe36b08` Source locator `/tasks/cli/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Legal Research Bench / weighted pass constitutional civil rights](https://benchize.com/benchmarks/vals-ai%3Alegal_research%3Aweighted_pass__constitutional_civil_rights.md) | 88.14% | SE ±2.46 pp | 7 / 50 | $1.04 | 220.84source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 88.144 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 1.043129 \| USD \| per_test \| Lower is better \| \| `latency` \| 220.844 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.46 pp sourcePopulation 82 stderr 2.458 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:52.389Z Evidence SHA-256 `26c58dfd47d0d2dc86f33c6d056f30309e0ee63570e7cd752c593b385f9d862e` Source locator `/tasks/weighted_pass__constitutional_civil_rights/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Legal Research Bench / business commercial](https://benchize.com/benchmarks/vals-ai%3Alegal_research%3Abusiness_commercial.md) | 41.86% | SE ±7.61 pp | 13 / 50 | $1.73 | 315.57source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 41.86 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 1.733887 \| USD \| per_test \| Lower is better \| \| `latency` \| 315.567 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±7.61 pp sourcePopulation 82 stderr 7.612 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:52.389Z Evidence SHA-256 `26c58dfd47d0d2dc86f33c6d056f30309e0ee63570e7cd752c593b385f9d862e` Source locator `/tasks/business_commercial/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Code Migration / code quality](https://benchize.com/benchmarks/vals-ai%3Acode-migration%3Acode_quality.md) | 3.74% | SE ±0.05 pp | 26 / 47 | $18.49 | 9,443.32source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_code_migration.Dttzoqk3.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 3.742 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 18.491664 \| USD \| per_test \| Lower is better \| \| `latency` \| 9443.316 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.05 pp sourcePopulation 79 stderr 0.053 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_code_migration.Dttzoqk3.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:46:12.712Z Evidence SHA-256 `99194597f2c76c402195be47839409046e18bd30454616da6dd40b736a088fa4` Source locator `/tasks/code_quality/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Terminal-Bench Science 0.1 · mathematical](https://benchize.com/benchmarks/terminal-science%3Av0-1-eval%3Amathematical.md) | 0% | SE ±0 pp | 9 / 9 | $510.98 | — | [terminal science](https://www.terminal-bench-science.ai/api/leaderboard?package=terminal-bench-science%2Fterminal-bench-science&name=v0-1-eval) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 0 \| percent \| reported \| Higher is better \| \| `total_cost` \| 510.9776022 \| USD \| all_trials \| Lower is better \| \| `total_tokens` \| 5225589584 \| tokens \| all_trials \| Lower is better \| ### Configuration agent mini-SWE-agent benchmarkVersion v0-1-eval domain mathematical Reasoning effort high ### Reported statistics Uncertainty Standard error: ±0 pp accuracyStandardError 0 passes 0 rank 9 tasks 51 ### Source [Original source](https://www.terminal-bench-science.ai/api/leaderboard?package=terminal-bench-science%2Fterminal-bench-science&name=v0-1-eval) Source terminal science Original model label Gemini 3.8 Flash Evaluated 2026-09-16T13:45:27.369938+00:00 Observed 2026-10-21T16:32:46.684Z Evidence SHA-256 `86e6278ca64da88b69c1efd6059a33248bd300f27673098f60184286b470c4a7` Source locator `/rows/8/metrics/domain_metrics/mathematical` |  |  |  |  |  |  |  |
| [SAGE / overall](https://benchize.com/benchmarks/vals-ai%3Asage%3Aoverall.md) | 35.06% | SE ±3.36 pp | 38 / 41 | $0.00895 | 28.64source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_sage.RRwIIzO5.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 35.064 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.008954 \| USD \| per_test \| Lower is better \| \| `latency` \| 28.639 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 30000 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±3.36 pp sourcePopulation 91 stderr 3.364 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_sage.RRwIIzO5.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:46:07.515Z Evidence SHA-256 `cc5fb3fecdb0382d492ff167a42ad84b9c5c3883373c78ed87baaa972b8b6f7c` Source locator `/tasks/overall/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Code Migration / cli](https://benchize.com/benchmarks/vals-ai%3Acode-migration%3Acli.md) | 25.87% | SE ±2.46 pp | 28 / 47 | $22.87 | 10,040.43source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_code_migration.Dttzoqk3.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 25.871 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 22.86835 \| USD \| per_test \| Lower is better \| \| `latency` \| 10040.43 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.46 pp sourcePopulation 79 stderr 2.456 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_code_migration.Dttzoqk3.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:46:12.712Z Evidence SHA-256 `99194597f2c76c402195be47839409046e18bd30454616da6dd40b736a088fa4` Source locator `/tasks/cli/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [BioMysteryBench / human solvable](https://benchize.com/benchmarks/vals-ai%3Abiomysterybench%3Ahuman_solvable.md) | 71.69% | SE ±1.83 pp | 16 / 20 | $1.23 | 267.24source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_biomysterybench.Bley-WAP.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 71.689 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 1.233913 \| USD \| per_test \| Lower is better \| \| `latency` \| 267.243 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Terminus 2 maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.83 pp sourcePopulation 27 stderr 1.826 tokenTotals {"cache_read_tokens":203046480,"cache_write_tokens":0,"input_tokens":228290259,"output_tokens":2626201,"reasoning_tokens":1325930} ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_biomysterybench.Bley-WAP.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:46:04.029Z Evidence SHA-256 `59bf79d38ec415a34ffb1c2901a16ecbbe101b7fa864e86104d2571fc7cfc23d` Source locator `/tasks/human_solvable/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Arena text / industry mathematical](https://benchize.com/benchmarks/lmarena%3Atext%3Aindustry_mathematical.md) | 1,518.56 | — | 6 / 148 | — | — | [Arena](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text/latest) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `rating` \| 1518.5580266909644 \| rating \| reported \| Higher is better \| ### Configuration category industry_mathematical config text Reasoning effort high thinkingEnabled true ### Reported statistics lower 1506.068965458837 rank 6 upper 1531.047087923092 variance 40.603494073547395 votes 2495 ### Source [Original source](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text/latest) Source Arena Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:40:19.533Z Evidence SHA-256 `8e027ef69b8de06c4b2a8b217bfe816b1f1536269bb257fd295b4ef1ba8246ac` Source locator `/rows/5918/row` |  |  |  |  |  |  |  |
| [Legal Research Bench / weighted pass civil litigation](https://benchize.com/benchmarks/vals-ai%3Alegal_research%3Aweighted_pass__civil_litigation.md) | 91.54% | SE ±2.33 pp | 4 / 50 | $1.36 | 254.44source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 91.539 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 1.35934 \| USD \| per_test \| Lower is better \| \| `latency` \| 254.44 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.33 pp sourcePopulation 82 stderr 2.326 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:52.389Z Evidence SHA-256 `26c58dfd47d0d2dc86f33c6d056f30309e0ee63570e7cd752c593b385f9d862e` Source locator `/tasks/weighted_pass__civil_litigation/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [CUA-bench / superhot](https://benchize.com/benchmarks/vals-ai%3Acua_bench%3Asuperhot.md) | 0% | SE ±0 pp | 8 / 8 | $51.31 | 34.29source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_cua_bench.DpgrKYWA.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 0 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 51.308013 \| USD \| per_test \| Lower is better \| \| `latency` \| 34.29 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort high harness google-computer-use maxOutputTokens 128000 provider google reasoning true Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 9 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_cua_bench.DpgrKYWA.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:46:16.228Z Evidence SHA-256 `8c415c63f28cd80496a4ccca27d1062ec98bb55268e58cba105b8cf2d7ee53d3` Source locator `/tasks/superhot/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [EMB / operating](https://benchize.com/benchmarks/vals-ai%3Aemb%3Aoperating.md) | 69.06% | SE ±5.61 pp | 12 / 47 | $6.43 | 621.68source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_emb.BY1LXsAw.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 69.063 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 6.429094 \| USD \| per_test \| Lower is better \| \| `latency` \| 621.677 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±5.61 pp sourcePopulation 76 stderr 5.613 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_emb.BY1LXsAw.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:57.495Z Evidence SHA-256 `d807aee620ea47152a1fca0257a8dd12a1be4512c76cc1744c1ae9e285180918` Source locator `/tasks/operating/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [LegalBench / overall](https://benchize.com/benchmarks/vals-ai%3Alegal_bench%3Aoverall.md) | 86.99% | SE ±0.43 pp | 6 / 44 | $0.00811 | 3.32source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 86.993 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.008109 \| USD \| per_test \| Lower is better \| \| `latency` \| 3.32 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.43 pp sourcePopulation 151 stderr 0.432 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/overall/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [ProgramBench / partial](https://benchize.com/benchmarks/vals-ai%3Aprogrambench%3Apartial.md) | 71.9% | SE ±1.91 pp | 16 / 40 | $10.93 | 2,799.27source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_programbench.C_J7ZU4X.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 71.899 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 10.928509 \| USD \| per_test \| Lower is better \| \| `latency` \| 2799.268 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.91 pp sourcePopulation 64 stderr 1.905 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_programbench.C_J7ZU4X.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:46:10.137Z Evidence SHA-256 `e9bbb4378c7a058441a21d456baf9316b96c2cf5b58d371a3c7c803f5000bd8c` Source locator `/tasks/partial/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Arena text style control / english](https://benchize.com/benchmarks/lmarena%3Atext_style_control%3Aenglish.md) | 1,497.89 | — | 11 / 151 | — | — | [Arena](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text_style_control/latest) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `rating` \| 1497.8879418820127 \| rating \| reported \| Higher is better \| ### Configuration category english config text_style_control Reasoning effort high thinkingEnabled true ### Reported statistics lower 1491.8301200227013 rank 11 upper 1503.9457637413238 variance 9.552934807334253 votes 16240 ### Source [Original source](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text_style_control/latest) Source Arena Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:40:19.533Z Evidence SHA-256 `d4875c03cb7ec0c6e8943c3145fd261225e8841b937240dc62a3032c4a798d84` Source locator `/rows/1647/row` |  |  |  |  |  |  |  |
| [HLAB-Vals-1.1 / criteria pass rate employment labor](https://benchize.com/benchmarks/vals-ai%3Ahlab%3Acriteria_pass_rate__employment_labor.md) | 88.58% | SE ±1.95 pp | 6 / 14 | $3.17 | 1,108.25source duration | [Vals AI · 1.1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 88.577 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 3.168674 \| USD \| per_test \| Lower is better \| \| `latency` \| 1108.248 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.95 pp sourcePopulation 24 stderr 1.95 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1.1 Observed 2026-10-21T16:45:52.890Z Evidence SHA-256 `14681f4a2310e0f257df9ec07c9c81e5dadc8f6f71634f02e174a02a8096d868` Source locator `/tasks/criteria_pass_rate__employment_labor/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Arena vision / overall (config: vision)](https://benchize.com/benchmarks/benchmark%3Avision%3Aoverall--config-vision.md) | 1,315.38 | — | 7 / 82 | — | — | [Arena](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/vision/latest) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `rating` \| 1315.379362654793 \| rating \| reported \| Higher is better \| ### Configuration category overall config vision Reasoning effort high thinkingEnabled true ### Reported statistics lower 1305.640678093509 rank 7 upper 1325.1180472160775 variance 24.689052105225585 votes 4645 ### Source [Original source](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/vision/latest) Source Arena Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:40:19.533Z Evidence SHA-256 `06e61305486dce3fec5dc776ac2fd8321104e84c346cf9e08fa6f8e8a66ce0d7` Source locator `/rows/6/row` |  |  |  |  |  |  |  |
| [Terminal-Bench Science 0.1 · earth](https://benchize.com/benchmarks/terminal-science%3Av0-1-eval%3Aearth.md) | 12.5% | SE ±6.75 pp | 3 / 9 | $117.58 | — | [terminal science](https://www.terminal-bench-science.ai/api/leaderboard?package=terminal-bench-science%2Fterminal-bench-science&name=v0-1-eval) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 12.5 \| percent \| reported \| Higher is better \| \| `total_cost` \| 117.57938257500001 \| USD \| all_trials \| Lower is better \| \| `total_tokens` \| 1055679015 \| tokens \| all_trials \| Lower is better \| ### Configuration agent mini-SWE-agent benchmarkVersion v0-1-eval domain earth Reasoning effort high ### Reported statistics Uncertainty Standard error: ±6.75 pp accuracyStandardError 6.750771560841522 passes 3 rank 9 tasks 24 ### Source [Original source](https://www.terminal-bench-science.ai/api/leaderboard?package=terminal-bench-science%2Fterminal-bench-science&name=v0-1-eval) Source terminal science Original model label Gemini 3.8 Flash Evaluated 2026-09-16T13:45:27.369938+00:00 Observed 2026-10-21T16:32:46.684Z Evidence SHA-256 `86e6278ca64da88b69c1efd6059a33248bd300f27673098f60184286b470c4a7` Source locator `/rows/8/metrics/domain_metrics/earth` |  |  |  |  |  |  |  |
| [EMB v1.1 / dcf](https://benchize.com/benchmarks/vals-ai%3Aemb-v1-1%3Adcf.md) | 66.12% | SE ±6.89 pp | 5 / 19 | $6.77 | 673.25source duration | [Vals AI · 1.1](https://www.vals.ai/_astro/benchmark_view_emb_v1_1.11L06J7L.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 66.115 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 6.77049 \| USD \| per_test \| Lower is better \| \| `latency` \| 673.252 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±6.89 pp sourcePopulation 30 stderr 6.886 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_emb_v1_1.11L06J7L.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1.1 Observed 2026-10-21T16:45:55.456Z Evidence SHA-256 `eea2ef2c83a6dc7d553c7c6531c287710ea7ffbc56ee1ccec06c85b104b2408e` Source locator `/tasks/dcf/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [EMB / lbo](https://benchize.com/benchmarks/vals-ai%3Aemb%3Albo.md) | 58.56% | SE ±6.15 pp | 18 / 47 | $13.51 | 1,125.23source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_emb.BY1LXsAw.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 58.56 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 13.510599 \| USD \| per_test \| Lower is better \| \| `latency` \| 1125.234 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±6.15 pp sourcePopulation 76 stderr 6.149 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_emb.BY1LXsAw.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:57.495Z Evidence SHA-256 `d807aee620ea47152a1fca0257a8dd12a1be4512c76cc1744c1ae9e285180918` Source locator `/tasks/lbo/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Legal Research Bench / health](https://benchize.com/benchmarks/vals-ai%3Alegal_research%3Ahealth.md) | 56.67% | SE ±9.2 pp | 22 / 50 | $1.9 | 282.9source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 56.667 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 1.899416 \| USD \| per_test \| Lower is better \| \| `latency` \| 282.896 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±9.2 pp sourcePopulation 82 stderr 9.202 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:52.389Z Evidence SHA-256 `26c58dfd47d0d2dc86f33c6d056f30309e0ee63570e7cd752c593b385f9d862e` Source locator `/tasks/health/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [HLAB-Vals-1.1 / white collar defense investigations](https://benchize.com/benchmarks/vals-ai%3Ahlab%3Awhite_collar_defense_investigations.md) | 0% | SE ±0 pp | 2 / 14 | $3.12 | 1,076.18source duration | [Vals AI · 1.1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 0 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 3.121405 \| USD \| per_test \| Lower is better \| \| `latency` \| 1076.182 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 24 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1.1 Observed 2026-10-21T16:45:52.890Z Evidence SHA-256 `14681f4a2310e0f257df9ec07c9c81e5dadc8f6f71634f02e174a02a8096d868` Source locator `/tasks/white_collar_defense_investigations/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Arena text / longer query](https://benchize.com/benchmarks/lmarena%3Atext%3Alonger_query.md) | 1,506.91 | — | 9 / 150 | — | — | [Arena](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text/latest) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `rating` \| 1506.9133001237606 \| rating \| reported \| Higher is better \| ### Configuration category longer_query config text Reasoning effort high thinkingEnabled true ### Reported statistics lower 1500.8634092434472 rank 9 upper 1512.963191004074 variance 9.527937528973842 votes 19059 ### Source [Original source](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text/latest) Source Arena Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:40:19.533Z Evidence SHA-256 `8e027ef69b8de06c4b2a8b217bfe816b1f1536269bb257fd295b4ef1ba8246ac` Source locator `/rows/8526/row` |  |  |  |  |  |  |  |
| [Arena text style control / multi turn](https://benchize.com/benchmarks/lmarena%3Atext_style_control%3Amulti_turn.md) | 1,504.02 | — | 9 / 151 | — | — | [Arena](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text_style_control/latest) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `rating` \| 1504.0224362319254 \| rating \| reported \| Higher is better \| ### Configuration category multi_turn config text_style_control Reasoning effort high thinkingEnabled true ### Reported statistics lower 1495.567145832969 rank 9 upper 1512.477726630882 variance 18.610621398713853 votes 6188 ### Source [Original source](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text_style_control/latest) Source Arena Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:40:19.533Z Evidence SHA-256 `d4875c03cb7ec0c6e8943c3145fd261225e8841b937240dc62a3032c4a798d84` Source locator `/rows/9324/row` |  |  |  |  |  |  |  |
| [ARC-AGI-3](https://benchize.com/benchmarks/arc-prize%3Av3_Semi_Private.md) | 10.37% | — | 14 / 53 | 4,398.28usd | — | [arc prize · v3](https://arcprize.org/media/data/leaderboard/v3.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 0.10366279223923121 \| score_fraction \| v3_Semi_Private \| Higher is better \| \| `total_cost` \| 4398.2756895 \| usd \| v3_Semi_Private \| Lower is better \| ### Configuration evaluationSplit v3_Semi_Private harness standard modelGroup google-gemini-3-8-flash modelType CoT Reasoning effort high sourceModelVersion google-gemini-3-8-flash-high thinkingEnabled true ### Source [Original source](https://arcprize.org/media/data/leaderboard/v3.json) Source arc prize Original model label Gemini 3.8 Flash (High) Source version v3 Observed 2026-10-21T16:34:21.112Z Evidence SHA-256 `1aaee2199f8c6d9b5a589aab2fd2480a729e9132f8cf8fbf7b49f81dbd5b2c49` Source locator `/evaluations/51` |  |  |  |  |  |  |  |
| [HLAB-Vals-1.1 / criteria pass rate real estate](https://benchize.com/benchmarks/vals-ai%3Ahlab%3Acriteria_pass_rate__real_estate.md) | 95.63% | SE ±1.11 pp | 4 / 14 | $4.56 | 1,120.3source duration | [Vals AI · 1.1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 95.627 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 4.5574 \| USD \| per_test \| Lower is better \| \| `latency` \| 1120.304 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.11 pp sourcePopulation 24 stderr 1.106 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1.1 Observed 2026-10-21T16:45:52.890Z Evidence SHA-256 `14681f4a2310e0f257df9ec07c9c81e5dadc8f6f71634f02e174a02a8096d868` Source locator `/tasks/criteria_pass_rate__real_estate/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [CyberBench v1.1 / overall](https://benchize.com/benchmarks/benchmark%3Acyber%3Aoverall.md) | 43.75% | SE ±2.21 pp | 25 / 28 | $1.31 | 588.85source duration | [Vals AI · 1.1](https://www.vals.ai/_astro/benchmark_view_cyberbench.DS6UOlu0.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 43.75 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 1.305189 \| USD \| per_test \| Lower is better \| \| `latency` \| 588.848 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.21 pp sourcePopulation 47 stderr 2.21 tokenTotals {"cache_read_tokens":450352484,"cache_write_tokens":null,"input_tokens":470052820,"output_tokens":1962374,"reasoning_tokens":1730282} ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_cyberbench.DS6UOlu0.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1.1 Observed 2026-10-21T16:46:15.238Z Evidence SHA-256 `943f593f929f8971a6b746269a2851b5febc2aad4e2f1066895f715033cdc7f1` Source locator `/tasks/overall/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Humanity's Sixth Sense](https://benchize.com/benchmarks/scale%3Ahss.md) | 41.6source score | CI 37.9–45.3 | 5 / 24 | — | — | [scale](https://labs.scale.com/leaderboard/hss) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 41.6 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort high Model configuration Gemini-3.8-Flash (high) sourceModelVersion Not reported thinkingEnabled true ### Reported statistics Uncertainty Confidence interval: 37.9–45.3 contaminationWarning Not reported rank 6 reportedConfidenceHalfWidth 3.7 ### Source [Original source](https://labs.scale.com/leaderboard/hss) Source scale Original model label Gemini-3.8-Flash (high) Observed 2026-10-21T16:35:26.953Z Evidence SHA-256 `0b7d37655e8886c99f8325ed125eafaa4f89312cd44e1a783ac0c49b45db842b` Source locator `/rows/5` |  |  |  |  |  |  |  |
| [ARC-AGI-1](https://benchize.com/benchmarks/arc-prize%3Av1_Semi_Private.md) | 98.5% | — | 1 / 213 | 0.21usd per task | — | [arc prize · v1](https://arcprize.org/media/data/leaderboard/v1.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 0.985 \| score_fraction \| v1_Semi_Private \| Higher is better \| \| `cost_per_task` \| 0.20570699249999985 \| usd_per_task \| v1_Semi_Private \| Lower is better \| ### Configuration evaluationSplit v1_Semi_Private harness Not reported modelGroup google-gemini-3-8-flash modelType CoT Reasoning effort high sourceModelVersion google-gemini-3-8-flash-high thinkingEnabled true ### Source [Original source](https://arcprize.org/media/data/leaderboard/v1.json) Source arc prize Original model label Gemini 3.8 Flash (High) Source version v1 Observed 2026-10-21T16:34:20.378Z Evidence SHA-256 `302d8267c0522bf450ec1deeac4d060e2cf2fb444ba0a4269014d9535de0bd1f` Source locator `/evaluations/241` |  |  |  |  |  |  |  |
| [Tax Agent Bench / controversy precedence analysis](https://benchize.com/benchmarks/vals-ai%3Atax_agent_bench%3Acontroversy_precedence_analysis.md) | 6.67% | SE ±4.63 pp | 23 / 46 | $0.92 | 207.33source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_tax_agent_bench.BFpexnpS.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 6.667 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.923283 \| USD \| per_test \| Lower is better \| \| `latency` \| 207.327 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±4.63 pp sourcePopulation 73 stderr 4.632 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_tax_agent_bench.BFpexnpS.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:55.958Z Evidence SHA-256 `446db182b67efb7923b3187a4040ec3a3ea492b8ad699a2b6d708d714cc063ac` Source locator `/tasks/controversy_precedence_analysis/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Legal Research Bench / criminal](https://benchize.com/benchmarks/vals-ai%3Alegal_research%3Acriminal.md) | 40.35% | SE ±6.56 pp | 9 / 50 | $1.53 | 315.87source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 40.351 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 1.534611 \| USD \| per_test \| Lower is better \| \| `latency` \| 315.871 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±6.56 pp sourcePopulation 82 stderr 6.556 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:52.389Z Evidence SHA-256 `26c58dfd47d0d2dc86f33c6d056f30309e0ee63570e7cd752c593b385f9d862e` Source locator `/tasks/criminal/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Vals Index / tax agent bench](https://benchize.com/benchmarks/vals-ai%3Avals_index%3Atax_agent_bench.md) | 32.12% | SE ±3.37 pp | 8 / 14 | $0.86 | 178.33source duration | [Vals AI · 2.2](https://www.vals.ai/_astro/benchmark_view_vals_index.ye6C1wH2.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 32.124 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.857214 \| USD \| per_test \| Lower is better \| \| `latency` \| 178.329 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±3.37 pp sourcePopulation 23 stderr 3.37 tokenTotals {"cache_read_tokens":208172664,"cache_write_tokens":null,"input_tokens":262586790,"output_tokens":7012701,"reasoning_tokens":4789793} ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_vals_index.ye6C1wH2.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 2.2 Observed 2026-10-21T16:45:49.430Z Evidence SHA-256 `0e95aee94cd2bf34bc4b43a8f4bb16c7817de138613c21011873fea4e0786b45` Source locator `/tasks/tax_agent_bench/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Terminal-Bench-Science / earth sciences](https://benchize.com/benchmarks/vals-ai%3Aterminal-bench-science%3Aearth_sciences.md) | 0% | SE ±0 pp | 14 / 30 | — | — | [Vals AI · 0.1](https://www.vals.ai/_astro/benchmark_view_terminal_bench_science.o-HR8Nyt.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 0 \| percent \| reported \| Higher is better \| ### Configuration computeEffort Not reported harness Mini-SWE-agent maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 41 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_terminal_bench_science.o-HR8Nyt.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 0.1 Observed 2026-10-21T16:46:03.529Z Evidence SHA-256 `a30a8618deaecf018ea1c8c9f675576cdd9ab79742d05788067db49afcc5f91b` Source locator `/tasks/earth_sciences/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [HLAB-Vals-1.1 / banking finance](https://benchize.com/benchmarks/vals-ai%3Ahlab%3Abanking_finance.md) | 0% | SE ±0 pp | 7 / 14 | $2.51 | 1,130.54source duration | [Vals AI · 1.1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 0 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 2.511915 \| USD \| per_test \| Lower is better \| \| `latency` \| 1130.544 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 24 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1.1 Observed 2026-10-21T16:45:52.890Z Evidence SHA-256 `14681f4a2310e0f257df9ec07c9c81e5dadc8f6f71634f02e174a02a8096d868` Source locator `/tasks/banking_finance/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [MMLU Pro / other](https://benchize.com/benchmarks/vals-ai%3Ammlu_pro%3Aother.md) | 87.55% | SE ±1.09 pp | 6 / 40 | $0.02 | 7.51source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 87.554 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.019777 \| USD \| per_test \| Lower is better \| \| `latency` \| 7.512 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.09 pp sourcePopulation 138 stderr 1.086 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:46:06.503Z Evidence SHA-256 `dab025507f46c716dd8d145e90ec0a737d9c56219853f41e4b1cc3c5e8c222b4` Source locator `/tasks/other/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [ProgramBench / strict](https://benchize.com/benchmarks/vals-ai%3Aprogrambench%3Astrict.md) | 1% | SE ±0.71 pp | 12 / 40 | $10.93 | 2,799.27source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_programbench.C_J7ZU4X.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 1 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 10.928509 \| USD \| per_test \| Lower is better \| \| `latency` \| 2799.268 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.71 pp sourcePopulation 64 stderr 0.705 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_programbench.C_J7ZU4X.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:46:10.137Z Evidence SHA-256 `e9bbb4378c7a058441a21d456baf9316b96c2cf5b58d371a3c7c803f5000bd8c` Source locator `/tasks/strict/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Arena text / korean](https://benchize.com/benchmarks/lmarena%3Atext%3Akorean.md) | 1,464.33 | — | 7 / 124 | — | — | [Arena](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text/latest) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `rating` \| 1464.332775592294 \| rating \| reported \| Higher is better \| ### Configuration category korean config text Reasoning effort high thinkingEnabled true ### Reported statistics lower 1440.750705716737 rank 7 upper 1487.9148454678507 variance 144.76636235688258 votes 740 ### Source [Original source](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text/latest) Source Arena Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:40:19.533Z Evidence SHA-256 `8e027ef69b8de06c4b2a8b217bfe816b1f1536269bb257fd295b4ef1ba8246ac` Source locator `/rows/8240/row` |  |  |  |  |  |  |  |
| [TapTap Maker public](https://benchize.com/benchmarks/taptap-maker%3Apublic.md) | 79.81% | Range 77.45–81.62% | 9 / 26 | $0.87 | — | [taptap maker](https://maker.taptap.cn/leaderboard/data/editions/2026-10-18-v6-voxel.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `l2_pass_rate` \| 0.7981 \| score_fraction \| reported \| Higher is better \| \| `cost_per_case` \| 0.8738 \| USD \| per_case \| Lower is better \| ### Configuration datasetVersion 0.2.3 edition 2026-10-18-v6-voxel harness pi Reasoning effort high reasoningMode model-managed ### Reported statistics Uncertainty Reported interval: 77.45–81.62% interval [0.7745,0.8162] intervalMethod bootstrap-case rank 9 rankInterval [7,12] ### Source [Original source](https://maker.taptap.cn/leaderboard/data/editions/2026-10-18-v6-voxel.json) Source taptap maker Original model label gemini-3.8-flash Observed 2026-10-21T16:33:19.620Z Evidence SHA-256 `ad2d7f1adeae0548a4edc0b5e75d11449b443449ff76533d04bc1fb1b6d7035e` Source locator `/public/8` |  |  |  |  |  |  |  |
| [EMB / template](https://benchize.com/benchmarks/vals-ai%3Aemb%3Atemplate.md) | 75.58% | SE ±3.2 pp | 6 / 47 | $9.5 | 813.6source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_emb.BY1LXsAw.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 75.58 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 9.502756 \| USD \| per_test \| Lower is better \| \| `latency` \| 813.595 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±3.2 pp sourcePopulation 76 stderr 3.201 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_emb.BY1LXsAw.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:57.495Z Evidence SHA-256 `d807aee620ea47152a1fca0257a8dd12a1be4512c76cc1744c1ae9e285180918` Source locator `/tasks/template/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [HLAB-Vals-1.1 / antitrust competition](https://benchize.com/benchmarks/vals-ai%3Ahlab%3Aantitrust_competition.md) | 0% | SE ±0 pp | 8 / 14 | $3.87 | 890.9source duration | [Vals AI · 1.1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 0 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 3.871498 \| USD \| per_test \| Lower is better \| \| `latency` \| 890.898 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 24 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1.1 Observed 2026-10-21T16:45:52.890Z Evidence SHA-256 `14681f4a2310e0f257df9ec07c9c81e5dadc8f6f71634f02e174a02a8096d868` Source locator `/tasks/antitrust_competition/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Finance Agent (v2) / comparables](https://benchize.com/benchmarks/vals-ai%3Afabv2%3Acomparables.md) | 52% | SE ±0.16 pp | 1 / 50 | $2.47 | 276.76source duration | [Vals AI · 2](https://www.vals.ai/_astro/benchmark_view_finance_agent_v2.DypCOnwd.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 52.004 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 2.465927 \| USD \| per_test \| Lower is better \| \| `latency` \| 276.762 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.16 pp sourcePopulation 80 stderr 0.158 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_finance_agent_v2.DypCOnwd.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 2 Observed 2026-10-21T16:45:57.035Z Evidence SHA-256 `f731b20a26c126725d125519f2b5841aa45254656dba859a9570d7ce7e5d8191` Source locator `/tasks/comparables/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / criteria pass rate white collar defense investigations](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Acriteria_pass_rate__white_collar_defense_investigations.md) | 87.88% | SE ±1.9 pp | 9 / 49 | $3.12 | 2,041.39source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 87.879 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 3.121405 \| USD \| per_test \| Lower is better \| \| `latency` \| 2041.394 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.9 pp sourcePopulation 81 stderr 1.897 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__white_collar_defense_investigations/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [SAGE / linear algebra](https://benchize.com/benchmarks/vals-ai%3Asage%3Alinear_algebra.md) | 41.7% | SE ±5.52 pp | 39 / 41 | $0.01 | 27.38source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_sage.RRwIIzO5.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 41.7 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.011572 \| USD \| per_test \| Lower is better \| \| `latency` \| 27.381 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 30000 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±5.52 pp sourcePopulation 91 stderr 5.515 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_sage.RRwIIzO5.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:46:07.515Z Evidence SHA-256 `cc5fb3fecdb0382d492ff167a42ad84b9c5c3883373c78ed87baaa972b8b6f7c` Source locator `/tasks/linear_algebra/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [MortgageTax / numerical extraction](https://benchize.com/benchmarks/vals-ai%3Amortgage_tax%3Anumerical_extraction.md) | 70.59% | SE ±1.29 pp | 24 / 31 | $0.07 | 21.32source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_mortgage_tax.DCd-oO5Y.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 70.588 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.068848 \| USD \| per_test \| Lower is better \| \| `latency` \| 21.324 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.29 pp sourcePopulation 98 stderr 1.285 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mortgage_tax.DCd-oO5Y.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:58.500Z Evidence SHA-256 `e5c844561f97d119d66d70269e726ecfb68be79eae39acf14ead3e95422e648c` Source locator `/tasks/numerical_extraction/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Arena text / russian](https://benchize.com/benchmarks/lmarena%3Atext%3Arussian.md) | 1,513.66 | — | 6 / 150 | — | — | [Arena](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text/latest) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `rating` \| 1513.6581951340045 \| rating \| reported \| Higher is better \| ### Configuration category russian config text Reasoning effort high thinkingEnabled true ### Reported statistics lower 1504.7146644088043 rank 6 upper 1522.6017258592046 variance 20.82197039356704 votes 5347 ### Source [Original source](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text/latest) Source Arena Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:40:19.533Z Evidence SHA-256 `8e027ef69b8de06c4b2a8b217bfe816b1f1536269bb257fd295b4ef1ba8246ac` Source locator `/rows/10389/row` |  |  |  |  |  |  |  |
| [Finance Agent v2.1 / market analysis](https://benchize.com/benchmarks/vals-ai%3Afabv2-1%3Amarket_analysis.md) | 70.48% | SE ±6.6 pp | 3 / 18 | $1.36 | 409.9source duration | [Vals AI · 2.1](https://www.vals.ai/_astro/benchmark_view_finance_agent_v2_1.ZbMhy9xo.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 70.484 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 1.357636 \| USD \| per_test \| Lower is better \| \| `latency` \| 409.898 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±6.6 pp sourcePopulation 28 stderr 6.602 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_finance_agent_v2_1.ZbMhy9xo.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 2.1 Observed 2026-10-21T16:45:54.954Z Evidence SHA-256 `f1b0c2d7dca4e73b52136862a9be059f7c8481c69f15d705ce90bf3bd9f78429` Source locator `/tasks/market_analysis/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [SWE-bench Multimodal / overall](https://benchize.com/benchmarks/vals-ai%3Aswebench-multimodal%3Aoverall.md) | 40.83% | SE ±2.25 pp | 11 / 14 | $3.69 | 1,338.3source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_swebench_multimodal.C7UOLrm7.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 40.833 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 3.688847 \| USD \| per_test \| Lower is better \| \| `latency` \| 1338.303 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.25 pp sourcePopulation 20 stderr 2.246 tokenTotals {"cache_read_tokens":5816508549,"cache_write_tokens":null,"input_tokens":6218542064,"output_tokens":38857502,"reasoning_tokens":34183849} ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_swebench_multimodal.C7UOLrm7.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:46:12.211Z Evidence SHA-256 `5302c42065b017d7deaa43d9e34be1ec69893b7ec4cceb6c07bfbabbd20b1f82` Source locator `/tasks/overall/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [HLAB-Vals-1.1 / criteria pass rate data privacy cybersecurity](https://benchize.com/benchmarks/vals-ai%3Ahlab%3Acriteria_pass_rate__data_privacy_cybersecurity.md) | 97.53% | SE ±0.93 pp | 4 / 14 | $2.44 | 1,295.87source duration | [Vals AI · 1.1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 97.527 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 2.439638 \| USD \| per_test \| Lower is better \| \| `latency` \| 1295.87 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.93 pp sourcePopulation 24 stderr 0.925 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark.CFXFrRQl.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1.1 Observed 2026-10-21T16:45:52.890Z Evidence SHA-256 `14681f4a2310e0f257df9ec07c9c81e5dadc8f6f71634f02e174a02a8096d868` Source locator `/tasks/criteria_pass_rate__data_privacy_cybersecurity/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [EMB / scratch](https://benchize.com/benchmarks/vals-ai%3Aemb%3Ascratch.md) | 68.39% | SE ±2.73 pp | 13 / 47 | $6.98 | 724.8source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_emb.BY1LXsAw.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 68.39 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 6.984969 \| USD \| per_test \| Lower is better \| \| `latency` \| 724.799 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.73 pp sourcePopulation 76 stderr 2.731 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_emb.BY1LXsAw.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:57.495Z Evidence SHA-256 `d807aee620ea47152a1fca0257a8dd12a1be4512c76cc1744c1ae9e285180918` Source locator `/tasks/scratch/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Arena text style control / korean](https://benchize.com/benchmarks/lmarena%3Atext_style_control%3Akorean.md) | 1,452.78 | — | 16 / 124 | — | — | [Arena](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text_style_control/latest) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `rating` \| 1452.7830023986485 \| rating \| reported \| Higher is better \| ### Configuration category korean config text_style_control Reasoning effort high thinkingEnabled true ### Reported statistics lower 1428.900188723122 rank 19 upper 1476.665816074175 variance 148.48233852910394 votes 740 ### Source [Original source](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text_style_control/latest) Source Arena Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:40:19.533Z Evidence SHA-256 `d4875c03cb7ec0c6e8943c3145fd261225e8841b937240dc62a3032c4a798d84` Source locator `/rows/8252/row` |  |  |  |  |  |  |  |
| [Arena text style control / exclude ties](https://benchize.com/benchmarks/lmarena%3Atext_style_control%3Aexclude_ties.md) | 1,514.07 | — | 9 / 151 | — | — | [Arena](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text_style_control/latest) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `rating` \| 1514.07277160875 \| rating \| reported \| Higher is better \| ### Configuration category exclude_ties config text_style_control Reasoning effort high thinkingEnabled true ### Reported statistics lower 1507.9594471059802 rank 9 upper 1520.1860961115199 variance 9.728787479078948 votes 30548 ### Source [Original source](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text_style_control/latest) Source Arena Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:40:19.533Z Evidence SHA-256 `d4875c03cb7ec0c6e8943c3145fd261225e8841b937240dc62a3032c4a798d84` Source locator `/rows/2062/row` |  |  |  |  |  |  |  |
| [Arena text style control / non english](https://benchize.com/benchmarks/lmarena%3Atext_style_control%3Anon_english.md) | 1,486.58 | — | 8 / 151 | — | — | [Arena](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text_style_control/latest) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `rating` \| 1486.5821041668896 \| rating \| reported \| Higher is better \| ### Configuration category non_english config text_style_control Reasoning effort high thinkingEnabled true ### Reported statistics lower 1481.1474609314123 rank 8 upper 1492.016747402367 variance 7.68857574050039 votes 25170 ### Source [Original source](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text_style_control/latest) Source Arena Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:40:19.533Z Evidence SHA-256 `d4875c03cb7ec0c6e8943c3145fd261225e8841b937240dc62a3032c4a798d84` Source locator `/rows/9738/row` |  |  |  |  |  |  |  |
| [Arena text style control / hard prompts](https://benchize.com/benchmarks/lmarena%3Atext_style_control%3Ahard_prompts.md) | 1,517.42 | — | 10 / 151 | — | — | [Arena](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text_style_control/latest) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `rating` \| 1517.4194658247275 \| rating \| reported \| Higher is better \| ### Configuration category hard_prompts config text_style_control Reasoning effort high thinkingEnabled true ### Reported statistics lower 1512.0119085370275 rank 10 upper 1522.8270231124277 variance 7.612127888039073 votes 26670 ### Source [Original source](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text_style_control/latest) Source Arena Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:40:19.533Z Evidence SHA-256 `d4875c03cb7ec0c6e8943c3145fd261225e8841b937240dc62a3032c4a798d84` Source locator `/rows/3460/row` |  |  |  |  |  |  |  |
| [Arena text / industry business and management and financial operations](https://benchize.com/benchmarks/lmarena%3Atext%3Aindustry_business_and_management_and_financial_operations.md) | 1,479.34 | — | 14 / 151 | — | — | [Arena](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text/latest) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `rating` \| 1479.3355746202355 \| rating \| reported \| Higher is better \| ### Configuration category industry_business_and_management_and_financial_operations config text Reasoning effort high thinkingEnabled true ### Reported statistics lower 1471.6505847259361 rank 14 upper 1487.0205645145352 variance 15.37412541228651 votes 7528 ### Source [Original source](https://datasets-server.huggingface.co/parquet/lmarena-ai%2Fleaderboard-dataset/text/latest) Source Arena Original model label Gemini 3.8 Flash (High) Observed 2026-10-21T16:40:19.533Z Evidence SHA-256 `8e027ef69b8de06c4b2a8b217bfe816b1f1536269bb257fd295b4ef1ba8246ac` Source locator `/rows/4296/row` |  |  |  |  |  |  |  |
| [Finance Agent (v2) / adjustments](https://benchize.com/benchmarks/vals-ai%3Afabv2%3Aadjustments.md) | 54.83% | SE ±2.4 pp | 6 / 50 | $1.83 | 203.65source duration | [Vals AI · 2](https://www.vals.ai/_astro/benchmark_view_finance_agent_v2.DypCOnwd.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 54.828 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 1.828394 \| USD \| per_test \| Lower is better \| \| `latency` \| 203.646 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.4 pp sourcePopulation 80 stderr 2.403 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_finance_agent_v2.DypCOnwd.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 2 Observed 2026-10-21T16:45:57.035Z Evidence SHA-256 `f731b20a26c126725d125519f2b5841aa45254656dba859a9570d7ce7e5d8191` Source locator `/tasks/adjustments/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [MMLU Pro / law](https://benchize.com/benchmarks/vals-ai%3Ammlu_pro%3Alaw.md) | 82.02% | SE ±1.16 pp | 6 / 40 | $0.03 | 10.67source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 82.016 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.030686 \| USD \| per_test \| Lower is better \| \| `latency` \| 10.668 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.16 pp sourcePopulation 138 stderr 1.162 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:46:06.503Z Evidence SHA-256 `dab025507f46c716dd8d145e90ec0a737d9c56219853f41e4b1cc3c5e8c222b4` Source locator `/tasks/law/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [SWE-bench Multimodal / lighthouse](https://benchize.com/benchmarks/vals-ai%3Aswebench-multimodal%3Alighthouse.md) | 20.37% | SE ±5.53 pp | 12 / 14 | — | — | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_swebench_multimodal.C7UOLrm7.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 20.37 \| percent \| reported \| Higher is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±5.53 pp sourcePopulation 20 stderr 5.532 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_swebench_multimodal.C7UOLrm7.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:46:12.211Z Evidence SHA-256 `5302c42065b017d7deaa43d9e34be1ec69893b7ec4cceb6c07bfbabbd20b1f82` Source locator `/tasks/lighthouse/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |
| [Finance Agent v2.1 / comparables](https://benchize.com/benchmarks/vals-ai%3Afabv2-1%3Acomparables.md) | 59.25% | SE ±6.48 pp | 3 / 18 | $2.58 | 543.81source duration | [Vals AI · 2.1](https://www.vals.ai/_astro/benchmark_view_finance_agent_v2_1.ZbMhy9xo.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 59.248 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 2.583678 \| USD \| per_test \| Lower is better \| \| `latency` \| 543.809 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±6.48 pp sourcePopulation 28 stderr 6.478 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_finance_agent_v2_1.ZbMhy9xo.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 2.1 Observed 2026-10-21T16:45:54.954Z Evidence SHA-256 `f1b0c2d7dca4e73b52136862a9be059f7c8481c69f15d705ce90bf3bd9f78429` Source locator `/tasks/comparables/google~1gemini-3.8-flash` |  |  |  |  |  |  |  |

[Next page](https://benchize.com/models/gemini-3-8-flash-high.md?cursor=3951123bbc7dd9d542101ec7489e9e207bcfddbc3076698ad2a4d78b6c6a4f72)

---
Cite as: Gemini 3.8 FlashHigh — Benchize, https://benchize.com/models/gemini-3-8-flash-high (retrieved 2026-10-21)