# Gemini 3.8 Flash

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

Google

Canonical URL: https://benchize.com/models/gemini-3-8-flash
Last updated: Wed, 21 Oct 2026 18:52:20 GMT

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

Google

## Benchmarks

| Benchmark | Score | ± | Rank | Cost | Source |  |
| --- | --- | --- | --- | --- | --- | --- |
| [DesignArena Overall frontend coding](https://benchize.com/benchmarks/designarena%3Acode.md) | 1,304elo | SE ±1.95 | 19 / 116 | — | [designarena · public](https://www.designarena.ai/api/leaderboard#board=code) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1304 \| elo \| code \| Higher is better \| ### Configuration arenaType models category allcategories inputModality text sourceModelVersion gemini-3.8-flash variationName public ### Reported statistics Uncertainty Standard error: ±1.95 averageGenerationTimeMs 122807 battles 46011 losses 22847 standardError 1.9482 winRate 50.3 wins 23164 ### Source [Original source](https://www.designarena.ai/api/leaderboard#board=code) Source designarena Original model label Gemini 3.8 Flash Source version public Observed 2026-10-21T16:37:57.301Z Evidence SHA-256 `0c897b58670f5e352bd123db999a40c0d1781a443964c502b90774a76a3a7758` Source locator `/data/22` |  |  |  |  |  |  |
| [DesignArena Native Android development](https://benchize.com/benchmarks/designarena%3Aandroid.md) | 1,071elo | — | 18 / 23 | — | [designarena · public](https://www.designarena.ai/api/leaderboard#board=android) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1071 \| elo \| android \| Higher is better \| ### Configuration arenaType agents category nativeapps inputModality Not reported sourceModelVersion gemini-3.8-flash variationName public ### Reported statistics averageGenerationTimeMs Not reported battles 329 losses 221 standardError Not reported winRate 32.8 wins 108 ### Source [Original source](https://www.designarena.ai/api/leaderboard#board=android) Source designarena Original model label Gemini 3.8 Flash Source version public Observed 2026-10-21T16:38:03.000Z Evidence SHA-256 `d9d0cd4bea54b696c28f907a98ba42ebf8dd4da9da201b6c70c95f76de35f525` Source locator `/data/29` |  |  |  |  |  |  |
| [Remote Labor Index (RLI)](https://benchize.com/benchmarks/scale%3Arli.md) | 5.83source score | CI 5.83–5.83 | 2 / 4 | — | [scale](https://labs.scale.com/leaderboard/rli) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 5.83 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration Gemini 3.8 Flash sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 5.83–5.83 contaminationWarning Not reported rank 7 reportedConfidenceHalfWidth 0 ### Source [Original source](https://labs.scale.com/leaderboard/rli) Source scale Original model label Gemini 3.8 Flash Observed 2026-10-21T16:35:35.546Z Evidence SHA-256 `d08c60517bef10b77b75d2b765ba3f22d2457693fb4bdf28a6e5d89e7e2b50a8` Source locator `/rows/6` |  |  |  |  |  |  |
| [ToneBench Scriptwriting](https://benchize.com/benchmarks/tonebench%3Ascripts.md) | 1,077.1 | Range 1,034.1–1,118.4 | 100 / 126 | — | [tonebench](https://benchmark.towardsai.com/data/leaderboard.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1077.1 \| rating \| reported \| Higher is better \| \| `overall` \| 69.13 \| percent \| reported \| Higher is better \| \| `anti_slop_numbers` \| 59.34 \| source_score \| rubric \| Higher is better \| \| `continuity_emotion` \| 73.45 \| source_score \| rubric \| Higher is better \| \| `cue_quality` \| 74.72 \| source_score \| rubric \| Higher is better \| \| `hook_strength` \| 74.09 \| source_score \| rubric \| Higher is better \| \| `length_adherence` \| 62 \| source_score \| rubric \| Higher is better \| \| `slop` \| 74.06 \| source_score \| rubric \| Higher is better \| \| `slop_score` \| 88.79 \| source_score \| rubric \| Higher is better \| \| `substance_accuracy` \| 71.72 \| source_score \| rubric \| Higher is better \| \| `tone_voice` \| 60.84 \| source_score \| rubric \| Higher is better \| \| `writing_quality` \| 67.98 \| source_score \| rubric \| Higher is better \| \| `youtube_best_practices` \| 71.9 \| source_score \| rubric \| Higher is better \| ### Configuration provider google Reasoning effort unknown scoreScope 10_task_current sourceConfiguration gemini-3-8-flash ### Reported statistics Uncertainty Reported interval: 1,034.1–1,118.4 interval [1034.1,1118.4] rank 105 runCount 5 ### Source [Original source](https://benchmark.towardsai.com/data/leaderboard.json) Source tonebench Original model label Gemini 3.8 Flash (medium thinking) Evaluated 2026-10-11 Observed 2026-10-21T16:32:48.862Z Evidence SHA-256 `4264338337f418d3ded7ace250cad64fe12cacbd4b79a312bfaf0ae9268a01c3` Source locator `/rows/104` |  |  |  |  |  |  |
| [BrokenArXiv 08/2026](https://benchize.com/benchmarks/matharena%3Aarxiv_false--august.md) | 27.38% | 95% CI 20.64–34.12% | 11 / 11 | 0.6usd per problem | [matharena · arxiv_false--august](https://matharena.ai/competition_tables/arxiv_false--august) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 27.38 \| percent \| competition \| Higher is better \| \| `cost_per_problem` \| 0.60 \| usd_per_problem \| competition \| Lower is better \| ### Configuration competition arxiv_false--august predicted false sourceModelName Gemini 3.8 Flash sourceModelVersion gemini/gemini-38-flash ### Reported statistics Uncertainty 95% confidence interval: 20.64–34.12% confidenceInterval95 6.74 problems 56 ### Source [Original source](https://matharena.ai/competition_tables/arxiv_false--august) Source matharena Original model label Gemini 3.8 Flash Source version arxiv_false--august Observed 2026-10-21T16:52:25.683Z Evidence SHA-256 `1facf024f36359a3fcf069fc3b747d7fbf6aeec67c7949ed674b50ec0447a209` Source locator `/table/rows/13` |  |  |  |  |  |  |
| [Longform Creative Writing](https://benchize.com/benchmarks/eq-bench%3Alongform-writing.md) | 76.8% | — | 23 / 93 | — | [eq bench · longform-writing](https://eqbench.com/creative_writing_longform.js) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `overall_score_100` \| 76.8 \| percent \| overall \| Higher is better \| \| `chapter1_avg` \| 16.48 \| rubric_score \| overall \| Higher is better \| \| `chapter2_avg` \| 16.09 \| rubric_score \| overall \| Higher is better \| \| `chapter3_avg` \| 16.01 \| rubric_score \| overall \| Higher is better \| \| `chapter4_avg` \| 15.67 \| rubric_score \| overall \| Higher is better \| \| `chapter5_avg` \| 16.04 \| rubric_score \| overall \| Higher is better \| \| `chapter6_avg` \| 15.85 \| rubric_score \| overall \| Higher is better \| \| `chapter7_avg` \| 15.91 \| rubric_score \| overall \| Higher is better \| \| `chapter8_avg` \| 16.06 \| rubric_score \| overall \| Higher is better \| \| `final_judgement_avg` \| 15.74 \| rubric_score \| overall \| Higher is better \| ### Configuration interactiveAdjustments none protocol longform-writing sourceModelVersion gemini-3.8-flash ### Source [Original source](https://eqbench.com/creative_writing_longform.js) Source eq bench Original model label gemini-3.8-flash Source version longform-writing Observed 2026-10-21T16:35:02.575Z Evidence SHA-256 `3ba283ce357ac7620e16db640a961879c30fe3ddc50a27de9a8ed4290784825c` Source locator `/leaderboardDataLongformV3/131` |  |  |  |  |  |  |
| [BrokenArXiv 04/2026](https://benchize.com/benchmarks/matharena%3Aarxiv_false--april.md) | 33.61% | 95% CI 25.23–41.99% | 6 / 17 | 0.15usd per problem | [matharena · arxiv_false--april](https://matharena.ai/competition_tables/arxiv_false--april) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 33.61 \| percent \| competition \| Higher is better \| \| `cost_per_problem` \| 0.15 \| usd_per_problem \| competition \| Lower is better \| ### Configuration competition arxiv_false--april predicted false sourceModelName Gemini 3.8 Flash sourceModelVersion gemini/gemini-38-flash ### Reported statistics Uncertainty 95% confidence interval: 25.23–41.99% confidenceInterval95 8.38 problems 61 ### Source [Original source](https://matharena.ai/competition_tables/arxiv_false--april) Source matharena Original model label Gemini 3.8 Flash Source version arxiv_false--april Observed 2026-10-21T16:52:24.402Z Evidence SHA-256 `e543f14dac0ba9663e8104a87738726c9848b89171a53fdca89b7baa4a892399` Source locator `/table/rows/7` |  |  |  |  |  |  |
| [Humanity’s Last Exam (HLE)](https://benchize.com/benchmarks/hle.md) | 44.52source score | CI 42.56–46.48 | 2 / 17 | — | [scale](https://labs.scale.com/leaderboard/humanitys_last_exam) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 44.52 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration Gemini 3.8 Flash sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 42.56–46.48 contaminationWarning Potential contamination warning: This model was evaluated after the public release of HLE, allowing model builder access to the prompts and solutions. rank 5 reportedConfidenceHalfWidth 1.96 ### Source [Original source](https://labs.scale.com/leaderboard/humanitys_last_exam) Source scale Original model label Gemini 3.8 Flash Observed 2026-10-21T16:35:32.694Z Evidence SHA-256 `9970bf9ee4f6b6b167633a94634adae7ac9c906447d6db087a8dc9877c1b71e0` Source locator `/rows/4` |  |  |  |  |  |  |
| [DesignArena Agentic web applications](https://benchize.com/benchmarks/designarena%3Awebapps.md) | 1,238elo | SE ±9.46 | 14 / 39 | — | [designarena · public](https://www.designarena.ai/api/leaderboard#board=webapps) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1238 \| elo \| webapps \| Higher is better \| ### Configuration arenaType agents category agon_webapps inputModality text sourceModelVersion gemini-3.8-flash variationName public ### Reported statistics Uncertainty Standard error: ±9.46 averageGenerationTimeMs Not reported battles 1500 losses 799 standardError 9.4556 winRate 46.7 wins 701 ### Source [Original source](https://www.designarena.ai/api/leaderboard#board=webapps) Source designarena Original model label Gemini 3.8 Flash Source version public Observed 2026-10-21T16:38:00.309Z Evidence SHA-256 `f955787274cf4ac627ada7cd33eaae2e08ef7297db92d3391340fcaa0b44a35f` Source locator `/data/16` |  |  |  |  |  |  |
| [Rapidata SVG generation — Preference](https://benchize.com/benchmarks/rapidata%3Asvg%3Arlb_1MFkkEMFnCNyKl.md) | 1,240.46elo | CI 1,214.71–1,266.2 | 8 / 37 | — | [rapidata · v1](https://api.rapidata.ai/benchmark/bmk_1MFkc3sXfbKCqf/standings/query?leaderboard_id%5Beq%5D=rlb_1MFkkEMFnCNyKl) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1240.4588189756976 \| elo \| rlb_1MFkkEMFnCNyKl \| Higher is better \| ### Configuration benchmarkId bmk_1MFkc3sXfbKCqf criterion rlb_1MFkkEMFnCNyKl participantId par_1U5GNxRa1BnOvM sourceModelVersion gemini-3.8-flash weighted false ### Reported statistics Uncertainty Confidence interval: 1,214.71–1,266.2 confidenceIntervalLowerOffset 25.744441846911286 confidenceIntervalUpperOffset 25.744441846911286 totalMatches 4522 wins 2910 ### Source [Original source](https://api.rapidata.ai/benchmark/bmk_1MFkc3sXfbKCqf/standings/query?leaderboard_id%5Beq%5D=rlb_1MFkkEMFnCNyKl) Source rapidata Original model label gemini-3.8-flash Source version v1 Observed 2026-10-21T16:37:40.980Z Evidence SHA-256 `b70f07a4924c82499996aba613ca8ec2d0aee9ec9ebd404e45c8bf250206cf40` Source locator `/items/12` |  |  |  |  |  |  |
| [ArXivMath 08/2026](https://benchize.com/benchmarks/matharena%3Aarxiv--august.md) | 40.35% | 95% CI 31.3–49.4% | 11 / 11 | 0.92usd per problem | [matharena · arxiv--august](https://matharena.ai/competition_tables/arxiv--august) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 40.35 \| percent \| competition \| Higher is better \| \| `cost_per_problem` \| 0.92 \| usd_per_problem \| competition \| Lower is better \| ### Configuration competition arxiv--august predicted false sourceModelName Gemini 3.8 Flash sourceModelVersion gemini/gemini-38-flash ### Reported statistics Uncertainty 95% confidence interval: 31.3–49.4% confidenceInterval95 9.05 problems 57 ### Source [Original source](https://matharena.ai/competition_tables/arxiv--august) Source matharena Original model label Gemini 3.8 Flash Source version arxiv--august Observed 2026-10-21T16:52:30.283Z Evidence SHA-256 `5924e5ecc15ff087ec1504b1e4ac44408b005c4f70e81e0d2184f930383a963c` Source locator `/table/rows/12` |  |  |  |  |  |  |
| [DesignArena SVG generation](https://benchize.com/benchmarks/designarena%3Asvgs.md) | 1,295elo | — | 10 / 77 | — | [designarena · public](https://www.designarena.ai/api/leaderboard#board=svgs) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1295 \| elo \| svgs \| Higher is better \| ### Configuration arenaType models category svg inputModality Not reported sourceModelVersion gemini-3.8-flash variationName public ### Reported statistics averageGenerationTimeMs 55508 battles 2235 losses 1010 standardError Not reported winRate 54.8 wins 1225 ### Source [Original source](https://www.designarena.ai/api/leaderboard#board=svgs) Source designarena Original model label Gemini 3.8 Flash Source version public Observed 2026-10-21T16:38:03.249Z Evidence SHA-256 `8cbc34a474a66ef5c64f3ffe557df9b55bf7e13368c07c0214223e3ed8f3f21e` Source locator `/data/13` |  |  |  |  |  |  |
| [BrokenArXiv 05/2026](https://benchize.com/benchmarks/matharena%3Aarxiv_false--may.md) | 19% | 95% CI 11.31–26.69% | 13 / 20 | 0.16usd per problem | [matharena · arxiv_false--may](https://matharena.ai/competition_tables/arxiv_false--may) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 19.00 \| percent \| competition \| Higher is better \| \| `cost_per_problem` \| 0.16 \| usd_per_problem \| competition \| Lower is better \| ### Configuration competition arxiv_false--may predicted false sourceModelName Gemini 3.8 Flash sourceModelVersion gemini/gemini-38-flash ### Reported statistics Uncertainty 95% confidence interval: 11.31–26.69% confidenceInterval95 7.69 problems 50 ### Source [Original source](https://matharena.ai/competition_tables/arxiv_false--may) Source matharena Original model label Gemini 3.8 Flash Source version arxiv_false--may Observed 2026-10-21T16:52:24.785Z Evidence SHA-256 `b317fca30109e8c78ed5d2b835fe227b8cfbee3b6f9cb6d7aaeadca866660969` Source locator `/table/rows/15` |  |  |  |  |  |  |
| [PRBench Finance](https://benchize.com/benchmarks/scale%3Aprbench-finance.md) | 49.51source score | CI 47.85–51.17 | 9 / 22 | — | [scale](https://labs.scale.com/leaderboard/prbench-finance) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 49.51 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration Gemini 3.8 Flash sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 47.85–51.17 contaminationWarning Not reported rank 12 reportedConfidenceHalfWidth 1.66 ### Source [Original source](https://labs.scale.com/leaderboard/prbench-finance) Source scale Original model label Gemini 3.8 Flash Observed 2026-10-21T16:35:34.798Z Evidence SHA-256 `948d67409f0ae9c93fd41428f634cfc93c7815a21f43690cecb9095c9d3ec8f1` Source locator `/rows/11` |  |  |  |  |  |  |
| [ArXivMath Overall](https://benchize.com/benchmarks/matharena%3Aoverall--arxivmath.md) | 52.57% | 95% CI 47.5–57.64% | 3 / 4 | 0.83usd per problem | [matharena · overall--arxivmath](https://matharena.ai/competition_tables/overall--arxivmath) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 52.57 \| percent \| competition \| Higher is better \| \| `cost_per_problem` \| 0.83 \| usd_per_problem \| competition \| Lower is better \| ### Configuration competition overall--arxivmath predicted false sourceModelName Gemini 3.8 Flash sourceModelVersion gemini/gemini-38-flash ### Reported statistics Uncertainty 95% confidence interval: 47.5–57.64% confidenceInterval95 5.07 problems Not reported ### Source [Original source](https://matharena.ai/competition_tables/overall--arxivmath) Source matharena Original model label Gemini 3.8 Flash Source version overall--arxivmath Observed 2026-10-21T16:52:26.434Z Evidence SHA-256 `6489e7ada75d40d19fd9a9486d8b138bdc4617d6e6af0c26df498de5697aafd9` Source locator `/table/rows/3` |  |  |  |  |  |  |
| [ArXivMath 05/2026](https://benchize.com/benchmarks/matharena%3Aarxiv--may.md) | 58.33% | 95% CI 49.51–67.15% | 11 / 21 | 0.15usd per problem | [matharena · arxiv--may](https://matharena.ai/competition_tables/arxiv--may) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 58.33 \| percent \| competition \| Higher is better \| \| `cost_per_problem` \| 0.15 \| usd_per_problem \| competition \| Lower is better \| ### Configuration competition arxiv--may predicted false sourceModelName Gemini 3.8 Flash sourceModelVersion gemini/gemini-38-flash ### Reported statistics Uncertainty 95% confidence interval: 49.51–67.15% confidenceInterval95 8.82 problems 40 ### Source [Original source](https://matharena.ai/competition_tables/arxiv--may) Source matharena Original model label Gemini 3.8 Flash Source version arxiv--may Observed 2026-10-21T16:52:29.369Z Evidence SHA-256 `758e1bba77e36f4ced2a51b4bc0d46bd956ed94ef6f57e93e1a974b75e76a4ce` Source locator `/table/rows/14` |  |  |  |  |  |  |
| [DesignArena Video to website](https://benchize.com/benchmarks/designarena%3Avideo-to-website.md) | 1,209elo | — | 5 / 10 | — | [designarena · public](https://www.designarena.ai/api/leaderboard#board=video-to-website) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1209 \| elo \| video-to-website \| Higher is better \| ### Configuration arenaType models category website inputModality video sourceModelVersion gemini-3.8-flash variationName public ### Reported statistics averageGenerationTimeMs 132359 battles 2584 losses 1248 standardError Not reported winRate 51.7 wins 1336 ### Source [Original source](https://www.designarena.ai/api/leaderboard#board=video-to-website) Source designarena Original model label Gemini 3.8 Flash Source version public Observed 2026-10-21T16:37:59.851Z Evidence SHA-256 `7497b4e107891bf38ad81750242b1b15b51d10f2bfc19781d85918d8f98cbbcb` Source locator `/data/5` |  |  |  |  |  |  |
| [DesignArena Godot game development](https://benchize.com/benchmarks/designarena%3Agodot-game-dev.md) | 1,273elo | SE ±14.5 | 5 / 14 | — | [designarena · public](https://www.designarena.ai/api/leaderboard#board=godot-game-dev) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1273 \| elo \| godot-game-dev \| Higher is better \| ### Configuration arenaType agents category agon_godot inputModality Not reported sourceModelVersion gemini-3.8-flash variationName public ### Reported statistics Uncertainty Standard error: ±14.5 averageGenerationTimeMs Not reported battles 664 losses 312 standardError 14.5043 winRate 53 wins 352 ### Source [Original source](https://www.designarena.ai/api/leaderboard#board=godot-game-dev) Source designarena Original model label Gemini 3.8 Flash Source version public Observed 2026-10-21T16:38:02.075Z Evidence SHA-256 `0aad65341ae5287f97ba93448674c557edad76a4e41056ec30cc32bf05d3a7e3` Source locator `/data/5` |  |  |  |  |  |  |
| [DesignArena Agentic presentation generation](https://benchize.com/benchmarks/designarena%3Aagentic-slides.md) | 1,177elo | SE ±12.75 | 14 / 19 | — | [designarena · public](https://www.designarena.ai/api/leaderboard#board=agentic-slides) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1177 \| elo \| agentic-slides \| Higher is better \| ### Configuration arenaType agents category agon_slides inputModality Not reported sourceModelVersion gemini-3.8-flash variationName public ### Reported statistics Uncertainty Standard error: ±12.75 averageGenerationTimeMs Not reported battles 822 losses 494 standardError 12.746 winRate 39.9 wins 328 ### Source [Original source](https://www.designarena.ai/api/leaderboard#board=agentic-slides) Source designarena Original model label Gemini 3.8 Flash Source version public Observed 2026-10-21T16:38:03.963Z Evidence SHA-256 `b732c5aba2e357cf6034f23084d5f95f8e8882f62b8bf3ee340f0cd5cd773724` Source locator `/data/20` |  |  |  |  |  |  |
| [ArXivLean 06/2026](https://benchize.com/benchmarks/matharena%3Aarxivlean--june.md) | 19.57% | 95% CI 8.11–31.03% | 6 / 10 | 1.4usd per problem | [matharena · arxivlean--june](https://matharena.ai/competition_tables/arxivlean--june) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 19.57 \| percent \| competition \| Higher is better \| \| `cost_per_problem` \| 1.40 \| usd_per_problem \| competition \| Lower is better \| ### Configuration competition arxivlean--june predicted false sourceModelName Gemini 3.8 Flash sourceModelVersion gemini/gemini-38-flash ### Reported statistics Uncertainty 95% confidence interval: 8.11–31.03% confidenceInterval95 11.46 problems 46 ### Source [Original source](https://matharena.ai/competition_tables/arxivlean--june) Source matharena Original model label Gemini 3.8 Flash Source version arxivlean--june Observed 2026-10-21T16:52:22.483Z Evidence SHA-256 `0869c2556b20f1879eb0ea99ea8151e805471cd9d6be57c652f764a775491b7d` Source locator `/table/rows/9` |  |  |  |  |  |  |
| [DesignArena Mobile application development](https://benchize.com/benchmarks/designarena%3Amobileapps.md) | 1,236elo | SE ±6.09 | 7 / 33 | — | [designarena · public](https://www.designarena.ai/api/leaderboard#board=mobileapps) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1236 \| elo \| mobileapps \| Higher is better \| ### Configuration arenaType agents category mobileapps inputModality Not reported sourceModelVersion gemini-3.8-flash variationName public ### Reported statistics Uncertainty Standard error: ±6.09 averageGenerationTimeMs Not reported battles 4238 losses 1931 standardError 6.0898 winRate 54.4 wins 2307 ### Source [Original source](https://www.designarena.ai/api/leaderboard#board=mobileapps) Source designarena Original model label Gemini 3.8 Flash Source version public Observed 2026-10-21T16:38:02.538Z Evidence SHA-256 `40af33d0e93e729486da375cbe954633e2458af0dd6ca6d9fbdcda90da650f7d` Source locator `/data/7` |  |  |  |  |  |  |
| [DesignArena ASCII art generation](https://benchize.com/benchmarks/designarena%3Aascii.md) | 1,264elo | — | 15 / 65 | — | [designarena · public](https://www.designarena.ai/api/leaderboard#board=ascii) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1264 \| elo \| ascii \| Higher is better \| ### Configuration arenaType models category ascii inputModality Not reported sourceModelVersion gemini-3.8-flash variationName public ### Reported statistics averageGenerationTimeMs 51712 battles 269 losses 131 standardError Not reported winRate 51.3 wins 138 ### Source [Original source](https://www.designarena.ai/api/leaderboard#board=ascii) Source designarena Original model label Gemini 3.8 Flash Source version public Observed 2026-10-21T16:38:03.513Z Evidence SHA-256 `0603b8c65b44ee887f5b566feb469465f59635137a6ac5c6d94c23bcf3776311` Source locator `/data/16` |  |  |  |  |  |  |
| [PRBench Legal](https://benchize.com/benchmarks/scale%3Aprbench-legal.md) | 48.9source score | CI 47.25–50.55 | 11 / 26 | — | [scale](https://labs.scale.com/leaderboard/prbench-legal) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 48.9 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration Gemini 3.8 Flash sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 47.25–50.55 contaminationWarning Not reported rank 15 reportedConfidenceHalfWidth 1.65 ### Source [Original source](https://labs.scale.com/leaderboard/prbench-legal) Source scale Original model label Gemini 3.8 Flash Observed 2026-10-21T16:35:35.197Z Evidence SHA-256 `9e2ca4e9350870cf508bf5fff603d0b795c81f7974a0ab980312bbbfa0785067` Source locator `/rows/14` |  |  |  |  |  |  |
| [SimpleBench](https://benchize.com/benchmarks/simplebench%3Aavg-at-5.md) | 82.4% | — | 5 / 62 | — | [simplebench](https://simple-bench.com/static/js/leaderboard-data.js) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 82.4 \| percent \| average_at_5 \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration Gemini 3.8 Flash thinkingEnabled Not reported ### Reported statistics rank 6th ### Source [Original source](https://simple-bench.com/static/js/leaderboard-data.js) Source simplebench Original model label Gemini 3.8 Flash Observed 2026-10-21T16:32:13.183Z Evidence SHA-256 `b2d0285ad3e4e15750ed20d39897e7558d1c05e8962e90c861b06f93c8019dc9` Source locator `/rows/7` |  |  |  |  |  |  |
| [ArXivMath 09/2026](https://benchize.com/benchmarks/matharena%3Aarxiv--september.md) | 50% | 95% CI 40.48–59.52% | 4 / 6 | 1.35usd per problem | [matharena · arxiv--september](https://matharena.ai/competition_tables/arxiv--september) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 50.00 \| percent \| competition \| Higher is better \| \| `cost_per_problem` \| 1.35 \| usd_per_problem \| competition \| Lower is better \| ### Configuration competition arxiv--september predicted false sourceModelName Gemini 3.8 Flash sourceModelVersion gemini/gemini-38-flash ### Reported statistics Uncertainty 95% confidence interval: 40.48–59.52% confidenceInterval95 9.52 problems 53 ### Source [Original source](https://matharena.ai/competition_tables/arxiv--september) Source matharena Original model label Gemini 3.8 Flash Source version arxiv--september Observed 2026-10-21T16:52:30.640Z Evidence SHA-256 `8973b7bb8ea675057eb4c03cedc79d40de49d0a1b15da24d387eaaad2d82b3be` Source locator `/table/rows/3` |  |  |  |  |  |  |
| [Rapidata Text-to-3D with LLM code — Coherence](https://benchize.com/benchmarks/rapidata%3Atext-to-3d-code%3Arlb_1YVpXufudyOLlc.md) | 912.76elo | CI 890.68–934.85 | 8 / 9 | — | [rapidata](https://api.rapidata.ai/benchmark/bmk_1YVpXgN69Mv8I9/standings/query?leaderboard_id%5Beq%5D=rlb_1YVpXufudyOLlc) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 912.7628219092358 \| elo \| rlb_1YVpXufudyOLlc \| Higher is better \| ### Configuration benchmarkId bmk_1YVpXgN69Mv8I9 criterion rlb_1YVpXufudyOLlc participantId par_1YVpXveBCmZ20L sourceModelVersion gemini-3.8-flash weighted false ### Reported statistics Uncertainty Confidence interval: 890.68–934.85 confidenceIntervalLowerOffset 22.08574857236431 confidenceIntervalUpperOffset 22.08574857236431 totalMatches 5383 wins 2404.5 ### Source [Original source](https://api.rapidata.ai/benchmark/bmk_1YVpXgN69Mv8I9/standings/query?leaderboard_id%5Beq%5D=rlb_1YVpXufudyOLlc) Source rapidata Original model label gemini-3.8-flash Source version Observed 2026-10-21T16:37:42.585Z Evidence SHA-256 `59243752590bb5f22389edbe87b8cc11b00e9a6fc49b5b94753544fc8780e58d` Source locator `/items/7` |  |  |  |  |  |  |
| [DesignArena 3D design generation](https://benchize.com/benchmarks/designarena%3A3d-design.md) | 1,295elo | — | 22 / 110 | — | [designarena · public](https://www.designarena.ai/api/leaderboard#board=3d-design) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1295 \| elo \| 3d-design \| Higher is better \| ### Configuration arenaType models category 3d inputModality Not reported sourceModelVersion gemini-3.8-flash variationName public ### Reported statistics averageGenerationTimeMs 79605 battles 6296 losses 3151 standardError Not reported winRate 50 wins 3145 ### Source [Original source](https://www.designarena.ai/api/leaderboard#board=3d-design) Source designarena Original model label Gemini 3.8 Flash Source version public Observed 2026-10-21T16:37:58.515Z Evidence SHA-256 `bcc0267f7bc7e9ec112306b3a35413cf78bbccbfaa838356232d5d5ed0b83d0f` Source locator `/data/26` |  |  |  |  |  |  |
| [ArXivMath 06/2026](https://benchize.com/benchmarks/matharena%3Aarxiv--june.md) | 67.36% | 95% CI 59.7–75.02% | 13 / 22 | 0.15usd per problem | [matharena · arxiv--june](https://matharena.ai/competition_tables/arxiv--june) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 67.36 \| percent \| competition \| Higher is better \| \| `cost_per_problem` \| 0.15 \| usd_per_problem \| competition \| Lower is better \| ### Configuration competition arxiv--june predicted false sourceModelName Gemini 3.8 Flash sourceModelVersion gemini/gemini-38-flash ### Reported statistics Uncertainty 95% confidence interval: 59.7–75.02% confidenceInterval95 7.66 problems 48 ### Source [Original source](https://matharena.ai/competition_tables/arxiv--june) Source matharena Original model label Gemini 3.8 Flash Source version arxiv--june Observed 2026-10-21T16:52:29.905Z Evidence SHA-256 `6832faff9be7e16a6c646c59017e30a2caedebe8cf02caf617a5ed9d8fb03bab` Source locator `/table/rows/16` |  |  |  |  |  |  |
| [DesignArena Browser game generation](https://benchize.com/benchmarks/designarena%3Agame-dev.md) | 1,305elo | — | 21 / 117 | — | [designarena · public](https://www.designarena.ai/api/leaderboard#board=game-dev) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1305 \| elo \| game-dev \| Higher is better \| ### Configuration arenaType models category gamedev inputModality Not reported sourceModelVersion gemini-3.8-flash variationName public ### Reported statistics averageGenerationTimeMs 106622 battles 12546 losses 6153 standardError Not reported winRate 51 wins 6393 ### Source [Original source](https://www.designarena.ai/api/leaderboard#board=game-dev) Source designarena Original model label Gemini 3.8 Flash Source version public Observed 2026-10-21T16:38:01.100Z Evidence SHA-256 `53db6a89112ce8cdf644568c666d4b8309aece764cfcedb291ffd52ae3934e30` Source locator `/data/25` |  |  |  |  |  |  |
| [DistressBench](https://benchize.com/benchmarks/scale%3Adistressbench.md) | 70.74source score | CI 69.51–71.97 | 14 / 27 | — | [scale](https://labs.scale.com/leaderboard/distressbench) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 70.74 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration gemini-3.8-flash sourceModelVersion gemini-3.8-flash thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 69.51–71.97 contaminationWarning Not reported rank 19 reportedConfidenceHalfWidth 1.23 ### Source [Original source](https://labs.scale.com/leaderboard/distressbench) Source scale Original model label gemini-3.8-flash Observed 2026-10-21T16:35:27.403Z Evidence SHA-256 `6ca6a76569e51fde037c8e2610bdf7bc39d99b54c69b6b4dc0dba9ec2a4e24da` Source locator `/rows/18` |  |  |  |  |  |  |
| [Humanity's Last Exam (Text Only)](https://benchize.com/benchmarks/scale%3Ahumanitys_last_exam_text_only.md) | 44.95source score | CI 42.85–47.05 | 2 / 25 | — | [scale](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 44.95 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration Gemini 3.8 Flash sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 42.85–47.05 contaminationWarning Potential contamination warning: This model was evaluated after the public release of HLE, allowing model builder access to the prompts and solutions. rank 6 reportedConfidenceHalfWidth 2.1 ### Source [Original source](https://labs.scale.com/leaderboard/humanitys_last_exam_text_only) Source scale Original model label Gemini 3.8 Flash Observed 2026-10-21T16:35:33.193Z Evidence SHA-256 `4004825ae58bdcc249c0d26dab811842c6d172b2a67bc35e69949fbf12ead757` Source locator `/rows/5` |  |  |  |  |  |  |
| [Rapidata SVG generation — Alignment](https://benchize.com/benchmarks/rapidata%3Asvg%3Arlb_1MFkk9yZHssERh.md) | 1,224.89elo | CI 1,199.53–1,250.26 | 8 / 37 | — | [rapidata · v1](https://api.rapidata.ai/benchmark/bmk_1MFkc3sXfbKCqf/standings/query?leaderboard_id%5Beq%5D=rlb_1MFkk9yZHssERh) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1224.8941774189616 \| elo \| rlb_1MFkk9yZHssERh \| Higher is better \| ### Configuration benchmarkId bmk_1MFkc3sXfbKCqf criterion rlb_1MFkk9yZHssERh participantId par_1U5GNxRa1BnOvM sourceModelVersion gemini-3.8-flash weighted false ### Reported statistics Uncertainty Confidence interval: 1,199.53–1,250.26 confidenceIntervalLowerOffset 25.36364175454264 confidenceIntervalUpperOffset 25.36364175454264 totalMatches 4455 wins 2823.5 ### Source [Original source](https://api.rapidata.ai/benchmark/bmk_1MFkc3sXfbKCqf/standings/query?leaderboard_id%5Beq%5D=rlb_1MFkk9yZHssERh) Source rapidata Original model label gemini-3.8-flash Source version v1 Observed 2026-10-21T16:37:40.342Z Evidence SHA-256 `2f5c8d08728735790b500d930d71b66f04253f1209f7c1c3da0db0846c163fcf` Source locator `/items/12` |  |  |  |  |  |  |
| [BrokenArXiv 06/2026](https://benchize.com/benchmarks/matharena%3Aarxiv_false--june.md) | 20.83% | 95% CI 13.17–28.49% | 14 / 21 | 0.16usd per problem | [matharena · arxiv_false--june](https://matharena.ai/competition_tables/arxiv_false--june) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 20.83 \| percent \| competition \| Higher is better \| \| `cost_per_problem` \| 0.16 \| usd_per_problem \| competition \| Lower is better \| ### Configuration competition arxiv_false--june predicted false sourceModelName Gemini 3.8 Flash sourceModelVersion gemini/gemini-38-flash ### Reported statistics Uncertainty 95% confidence interval: 13.17–28.49% confidenceInterval95 7.66 problems 54 ### Source [Original source](https://matharena.ai/competition_tables/arxiv_false--june) Source matharena Original model label Gemini 3.8 Flash Source version arxiv_false--june Observed 2026-10-21T16:52:25.323Z Evidence SHA-256 `f6d7e2280d05706643d46ae7786bd3fad6d912e4893781c3f62c73deebf38459` Source locator `/table/rows/20` |  |  |  |  |  |  |
| [Rapidata Text-to-3D with LLM code — Creativity](https://benchize.com/benchmarks/rapidata%3Atext-to-3d-code%3Arlb_1YVqTDVUylKrUp.md) | 1,330.58elo | CI 1,305.43–1,355.74 | 3 / 9 | — | [rapidata](https://api.rapidata.ai/benchmark/bmk_1YVpXgN69Mv8I9/standings/query?leaderboard_id%5Beq%5D=rlb_1YVqTDVUylKrUp) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1330.583641097671 \| elo \| rlb_1YVqTDVUylKrUp \| Higher is better \| ### Configuration benchmarkId bmk_1YVpXgN69Mv8I9 criterion rlb_1YVqTDVUylKrUp participantId par_1YVpXveBCmZ20L sourceModelVersion gemini-3.8-flash weighted false ### Reported statistics Uncertainty Confidence interval: 1,305.43–1,355.74 confidenceIntervalLowerOffset 25.15475155708691 confidenceIntervalUpperOffset 25.15475155708691 totalMatches 4817 wins 3265 ### Source [Original source](https://api.rapidata.ai/benchmark/bmk_1YVpXgN69Mv8I9/standings/query?leaderboard_id%5Beq%5D=rlb_1YVqTDVUylKrUp) Source rapidata Original model label gemini-3.8-flash Source version Observed 2026-10-21T16:37:42.177Z Evidence SHA-256 `4ce251bc655221de3fef117b85e05f775323a39098049de2bbfd4835ca2c6293` Source locator `/items/2` |  |  |  |  |  |  |
| [ArXivMath 04/2026](https://benchize.com/benchmarks/matharena%3Aarxiv--april.md) | 64.17% | 95% CI 55.59–72.75% | 5 / 19 | 0.15usd per problem | [matharena · arxiv--april](https://matharena.ai/competition_tables/arxiv--april) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 64.17 \| percent \| competition \| Higher is better \| \| `cost_per_problem` \| 0.15 \| usd_per_problem \| competition \| Lower is better \| ### Configuration competition arxiv--april predicted false sourceModelName Gemini 3.8 Flash sourceModelVersion gemini/gemini-38-flash ### Reported statistics Uncertainty 95% confidence interval: 55.59–72.75% confidenceInterval95 8.58 problems 40 ### Source [Original source](https://matharena.ai/competition_tables/arxiv--april) Source matharena Original model label Gemini 3.8 Flash Source version arxiv--april Observed 2026-10-21T16:52:28.840Z Evidence SHA-256 `5139c835884bfab38f18d23a248814af4680927310bc5cd2a86ea2eab779e0a7` Source locator `/table/rows/5` |  |  |  |  |  |  |
| [BrokenArXiv Overall](https://benchize.com/benchmarks/matharena%3Aoverall--brokenarxiv.md) | 24.98% | 95% CI 20.57–29.39% | 4 / 4 | 0.5usd per problem | [matharena · overall--brokenarxiv](https://matharena.ai/competition_tables/overall--brokenarxiv) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 24.98 \| percent \| competition \| Higher is better \| \| `cost_per_problem` \| 0.50 \| usd_per_problem \| competition \| Lower is better \| ### Configuration competition overall--brokenarxiv predicted false sourceModelName Gemini 3.8 Flash sourceModelVersion gemini/gemini-38-flash ### Reported statistics Uncertainty 95% confidence interval: 20.57–29.39% confidenceInterval95 4.41 problems Not reported ### Source [Original source](https://matharena.ai/competition_tables/overall--brokenarxiv) Source matharena Original model label Gemini 3.8 Flash Source version overall--brokenarxiv Observed 2026-10-21T16:52:23.171Z Evidence SHA-256 `8af403d03bad2e0ac76660844e9b21df38cc214b99fc352f6374309a5ea83df6` Source locator `/table/rows/5` |  |  |  |  |  |  |
| [Rapidata Text-to-3D with LLM code — Preference](https://benchize.com/benchmarks/rapidata%3Atext-to-3d-code%3Arlb_1YVpXugYHlm63x.md) | 1,186.73elo | CI 1,162.34–1,211.12 | 6 / 9 | — | [rapidata](https://api.rapidata.ai/benchmark/bmk_1YVpXgN69Mv8I9/standings/query?leaderboard_id%5Beq%5D=rlb_1YVpXugYHlm63x) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1186.7301290713165 \| elo \| rlb_1YVpXugYHlm63x \| Higher is better \| ### Configuration benchmarkId bmk_1YVpXgN69Mv8I9 criterion rlb_1YVpXugYHlm63x participantId par_1YVpXveBCmZ20L sourceModelVersion gemini-3.8-flash weighted false ### Reported statistics Uncertainty Confidence interval: 1,162.34–1,211.12 confidenceIntervalLowerOffset 24.390498639687166 confidenceIntervalUpperOffset 24.390498639687166 totalMatches 5383 wins 3185.5 ### Source [Original source](https://api.rapidata.ai/benchmark/bmk_1YVpXgN69Mv8I9/standings/query?leaderboard_id%5Beq%5D=rlb_1YVpXugYHlm63x) Source rapidata Original model label gemini-3.8-flash Source version Observed 2026-10-21T16:37:43.410Z Evidence SHA-256 `a30cb5e42eb353cd3fadbb4c6519873ad22d3f0528a4e569fbbe40160928d934` Source locator `/items/5` |  |  |  |  |  |  |
| [TubeLab Scriptwriting](https://benchize.com/benchmarks/tubelab%3Ascriptwriting.md) | 78.5% | — | 8 / 16 | 0.17ratio | [tubelab](https://tubelab.net/_next/static/chunks/pages/benchmark/scriptwriting-7ad7196ff3093ad4.js) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `quality` \| 78.5 \| percent \| source_reported \| Higher is better \| \| `craft` \| 78.5 \| percent \| source_reported \| Higher is better \| \| `relative_cost_per_script` \| 0.17 \| ratio \| source_reported \| Lower is better \| \| `script_time` \| 1.8 \| minutes \| source_reported \| Lower is better \| \| `hook` \| 8.3 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `retention` \| 7.8 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `storytelling` \| 7.4 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `voiceMatch` \| 7.4 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `pacing` \| 7.5 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `clarity` \| 8.6 \| score_out_of_10 \| craft_rubric \| Higher is better \| ### Configuration judgingProtocol blind_model_judges Reasoning effort unknown taskCount 12 ### Reported statistics confidence {"high":80.6,"low":76.2} headToHead {"losses":8,"ties":2,"winPercent":27,"wins":2} rank 12 ### Source [Original source](https://tubelab.net/_next/static/chunks/pages/benchmark/scriptwriting-7ad7196ff3093ad4.js) Source tubelab Original model label Gemini 3.8 Flash Observed 2026-10-21T16:34:03.856Z Evidence SHA-256 `72f9d867f388256af8fc6dca12e8c7b6f1ddcbb97f0712da339b73102686761f` Source locator `/rows/11` |  |  |  |  |  |  |
| [Humanity's Last Exam (Diamond)](https://benchize.com/benchmarks/scale%3Ahle-diamond.md) | 34.3source score | CI 31.4–37.2 | 7 / 12 | — | [scale](https://labs.scale.com/leaderboard/hle-diamond) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 34.3 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration gemini-3.8-flash sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 31.4–37.2 contaminationWarning Potential contamination warning: This model was evaluated after the public release of HLE, allowing model builder access to the prompts and solutions. rank 8 reportedConfidenceHalfWidth 2.9 ### Source [Original source](https://labs.scale.com/leaderboard/hle-diamond) Source scale Original model label gemini-3.8-flash Observed 2026-10-21T16:35:27.820Z Evidence SHA-256 `c60d42f5aacdf787d9ccc8b70817766d86707776a7f8b44ef0337cb9ae60bf20` Source locator `/rows/7` |  |  |  |  |  |  |
| [DesignArena Website generation](https://benchize.com/benchmarks/designarena%3Awebsite.md) | 1,301elo | — | 19 / 123 | — | [designarena · public](https://www.designarena.ai/api/leaderboard#board=website) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1301 \| elo \| website \| Higher is better \| ### Configuration arenaType models category website inputModality text sourceModelVersion gemini-3.8-flash variationName public ### Reported statistics averageGenerationTimeMs 136500 battles 29455 losses 14713 standardError Not reported winRate 50 wins 14742 ### Source [Original source](https://www.designarena.ai/api/leaderboard#board=website) Source designarena Original model label Gemini 3.8 Flash Source version public Observed 2026-10-21T16:37:57.757Z Evidence SHA-256 `51873e97776a3c5ae706b24cc2c60384793dd18398df970e213a8e7c276d8ac0` Source locator `/data/21` |  |  |  |  |  |  |
| [Rapidata Text-to-3D with LLM code — Overall preference](https://benchize.com/benchmarks/rapidata%3Atext-to-3d-code%3Aoverall.md) | 1,107.15elo | CI 1,095.58–1,118.72 | 6 / 9 | — | [rapidata](https://api.rapidata.ai/benchmark/bmk_1YVpXgN69Mv8I9/standings/query) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1107.1464986402145 \| elo \| overall \| Higher is better \| ### Configuration benchmarkId bmk_1YVpXgN69Mv8I9 criterion overall participantId par_1YVpXveBCmZ20L sourceModelVersion gemini-3.8-flash weighted false ### Reported statistics Uncertainty Confidence interval: 1,095.58–1,118.72 confidenceIntervalLowerOffset 11.56992243960849 confidenceIntervalUpperOffset 11.56992243960849 totalMatches 20966.20185184822 wins 11674.836566029646 ### Source [Original source](https://api.rapidata.ai/benchmark/bmk_1YVpXgN69Mv8I9/standings/query) Source rapidata Original model label gemini-3.8-flash Source version Observed 2026-10-21T16:37:41.757Z Evidence SHA-256 `048d6dec5a711f63fb70b09a6495cf440b0bf6f34f80156b97832c611e7a2913` Source locator `/items/5` |  |  |  |  |  |  |
| [DesignArena Full-stack application development](https://benchize.com/benchmarks/designarena%3Afullstack.md) | 1,244elo | SE ±10.76 | 10 / 31 | — | [designarena · public](https://www.designarena.ai/api/leaderboard#board=fullstack) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1244 \| elo \| fullstack \| Higher is better \| ### Configuration arenaType agents category fullstack inputModality Not reported sourceModelVersion gemini-3.8-flash variationName public ### Reported statistics Uncertainty Standard error: ±10.76 averageGenerationTimeMs Not reported battles 1155 losses 599 standardError 10.7607 winRate 48.1 wins 556 ### Source [Original source](https://www.designarena.ai/api/leaderboard#board=fullstack) Source designarena Original model label Gemini 3.8 Flash Source version public Observed 2026-10-21T16:38:00.857Z Evidence SHA-256 `aa171459dd44cfad57fba8da746c40b27832b57dc731b1198bdb71c73ddd0f97` Source locator `/data/13` |  |  |  |  |  |  |
| [Creative Writing v3](https://benchize.com/benchmarks/eq-bench%3Acreative-writing-v3.md) | 1,747.9elo | — | 27 / 90 | — | [eq bench · creative-writing-v3](https://eqbench.com/creative_writing.js) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo_score` \| 1747.9 \| elo \| overall \| Higher is better \| \| `creative_writing_score` \| 16.54 \| rubric_score \| overall \| Higher is better \| ### Configuration interactiveAdjustments none protocol creative-writing-v3 sourceModelVersion gemini-3.8-flash ### Source [Original source](https://eqbench.com/creative_writing.js) Source eq bench Original model label gemini-3.8-flash Source version creative-writing-v3 Observed 2026-10-21T16:35:01.421Z Evidence SHA-256 `722ddb132e8d3a5e5c39b8d445f8475f3f7e24893a007260c6f3e3cce5c1ee18` Source locator `/leaderboardDataCreativeWritingV3/29` |  |  |  |  |  |  |
| [Vending-Bench 2](https://benchize.com/benchmarks/vendingbench%3Av2.md) | — | — | — | $5,093.79 | [vendingbench](https://andonlabs.com/_app/immutable/chunks/Cjn0_qj9.js) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `final_balance` \| 5093.793333333332 \| USD \| one_simulated_year \| Higher is better \| \| `geometric_mean_balance` \| 4978.1163620343395 \| USD \| positive_final_balances \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration Gemini 3.8 Flash thinkingEnabled Not reported ### Reported statistics confidenceHalfWidth95 960.7212711800438 positiveBalanceRuns 6 runCount 6 standardDeviation 1200.651479300578 standardError 490.1639138673693 ### Source [Original source](https://andonlabs.com/_app/immutable/chunks/Cjn0_qj9.js) Source vendingbench Original model label Gemini 3.8 Flash Observed 2026-10-21T16:36:40.592Z Evidence SHA-256 `1a0b043147050298c405bcb743d73e425d532464202231cd18c27f836fa5f223` Source locator `/rows/45` |  |  |  |  |  |  |
| [Rapidata SVG generation — Overall preference](https://benchize.com/benchmarks/rapidata%3Asvg%3Aoverall.md) | 1,081.94elo | CI 1,067.97–1,095.92 | 10 / 37 | — | [rapidata · v1](https://api.rapidata.ai/benchmark/bmk_1MFkc3sXfbKCqf/standings/query) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1081.9448689266424 \| elo \| overall \| Higher is better \| ### Configuration benchmarkId bmk_1MFkc3sXfbKCqf criterion overall participantId par_1U5GNxRa1BnOvM sourceModelVersion gemini-3.8-flash weighted false ### Reported statistics Uncertainty Confidence interval: 1,067.97–1,095.92 confidenceIntervalLowerOffset 13.974713000856127 confidenceIntervalUpperOffset 13.974713000856127 totalMatches 13498.903874908177 wins 7499.696857743417 ### Source [Original source](https://api.rapidata.ai/benchmark/bmk_1MFkc3sXfbKCqf/standings/query) Source rapidata Original model label gemini-3.8-flash Source version v1 Observed 2026-10-21T16:37:39.078Z Evidence SHA-256 `4b73ff7e4ca0ecfd64f7888c2d64f3398a058c945c4644acd54eddb330e16d66` Source locator `/items/19` |  |  |  |  |  |  |
| [HiL-Bench (Human-in-Loop Benchmark)](https://benchize.com/benchmarks/scale%3Ahil.md) | 41.47source score | CI 35.81–47.13 | 11 / 22 | — | [scale](https://labs.scale.com/leaderboard/hil) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 41.47 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration Gemini 3.8 Flash sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 35.81–47.13 contaminationWarning Not reported rank 16 reportedConfidenceHalfWidth 5.66 ### Source [Original source](https://labs.scale.com/leaderboard/hil) Source scale Original model label Gemini 3.8 Flash Observed 2026-10-21T16:35:30.082Z Evidence SHA-256 `043a333e27453edada150667c7f4439b2aab463f37928c2153c6aa79f33dd1b2` Source locator `/rows/15` |  |  |  |  |  |  |
| [DesignArena UI component generation](https://benchize.com/benchmarks/designarena%3Aui-components.md) | 1,318elo | — | 16 / 112 | — | [designarena · public](https://www.designarena.ai/api/leaderboard#board=ui-components) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1318 \| elo \| ui-components \| Higher is better \| ### Configuration arenaType models category uicomponent inputModality Not reported sourceModelVersion gemini-3.8-flash variationName public ### Reported statistics averageGenerationTimeMs 88490 battles 1768 losses 816 standardError Not reported winRate 53.9 wins 952 ### Source [Original source](https://www.designarena.ai/api/leaderboard#board=ui-components) Source designarena Original model label Gemini 3.8 Flash Source version public Observed 2026-10-21T16:37:57.993Z Evidence SHA-256 `46482ec1b4012638eace65493daea4750e2088098018fd9d5fc46126fc1f8557` Source locator `/data/17` |  |  |  |  |  |  |
| [FrontierSWE v2 abs](https://benchize.com/benchmarks/frontierswe%3Av2%3Aabs.md) | 19.63source score | — | 11 / 15 | — | [frontierswe](https://www.frontierswe.com/) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `overall` \| 19.626902941176464 \| source_score \| mean \| Higher is better \| \| `overall_best` \| 31.642826470588243 \| source_score \| best \| Higher is better \| \| `overall_worst` \| 9.514232352941177 \| source_score \| worst \| Higher is better \| \| `implementation` \| 23.043864285714285 \| source_score \| mean \| Higher is better \| \| `performance` \| 14.876966666666666 \| source_score \| mean \| Higher is better \| \| `research` \| 16.47233333333334 \| source_score \| mean \| Higher is better \| ### Configuration aggregation mean harness proximus normalization abs Reasoning effort unknown ### Source [Original source](https://www.frontierswe.com/) Source frontierswe Original model label Gemini 3.8 Flash Observed 2026-10-21T16:31:27.602Z Evidence SHA-256 `818e027068cf61ff1aadd8281b6e49e9c6795221015210601b5e34d0d223e5b5` Source locator `/rows/14` |  |  |  |  |  |  |
| [Rapidata SVG generation — Coherence](https://benchize.com/benchmarks/rapidata%3Asvg%3Arlb_1MFkkBzpuU8fTG.md) | 817.4elo | CI 793.41–841.39 | 31 / 37 | — | [rapidata · v1](https://api.rapidata.ai/benchmark/bmk_1MFkc3sXfbKCqf/standings/query?leaderboard_id%5Beq%5D=rlb_1MFkkBzpuU8fTG) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 817.3998254093667 \| elo \| rlb_1MFkkBzpuU8fTG \| Higher is better \| ### Configuration benchmarkId bmk_1MFkc3sXfbKCqf criterion rlb_1MFkkBzpuU8fTG participantId par_1U5GNxRa1BnOvM sourceModelVersion gemini-3.8-flash weighted false ### Reported statistics Uncertainty Confidence interval: 793.41–841.39 confidenceIntervalLowerOffset 23.994557892282273 confidenceIntervalUpperOffset 23.994557892282273 totalMatches 4522 wins 1761 ### Source [Original source](https://api.rapidata.ai/benchmark/bmk_1MFkc3sXfbKCqf/standings/query?leaderboard_id%5Beq%5D=rlb_1MFkkBzpuU8fTG) Source rapidata Original model label gemini-3.8-flash Source version v1 Observed 2026-10-21T16:37:39.715Z Evidence SHA-256 `aa3cb94ff16f544c58adeea35b9087d281c785058b8ce1ae0f98a7ff84920930` Source locator `/items/58` |  |  |  |  |  |  |
| [Rapidata Text-to-3D with LLM code — Alignment](https://benchize.com/benchmarks/rapidata%3Atext-to-3d-code%3Arlb_1YVpXug2nDeUlV.md) | 1,052.44elo | CI 1,029.39–1,075.49 | 7 / 9 | — | [rapidata](https://api.rapidata.ai/benchmark/bmk_1YVpXgN69Mv8I9/standings/query?leaderboard_id%5Beq%5D=rlb_1YVpXug2nDeUlV) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1052.441636750989 \| elo \| rlb_1YVpXug2nDeUlV \| Higher is better \| ### Configuration benchmarkId bmk_1YVpXgN69Mv8I9 criterion rlb_1YVpXug2nDeUlV participantId par_1YVpXveBCmZ20L sourceModelVersion gemini-3.8-flash weighted false ### Reported statistics Uncertainty Confidence interval: 1,029.39–1,075.49 confidenceIntervalLowerOffset 23.04792287730902 confidenceIntervalUpperOffset 23.04792287730902 totalMatches 5383 wins 2820 ### Source [Original source](https://api.rapidata.ai/benchmark/bmk_1YVpXgN69Mv8I9/standings/query?leaderboard_id%5Beq%5D=rlb_1YVpXug2nDeUlV) Source rapidata Original model label gemini-3.8-flash Source version Observed 2026-10-21T16:37:42.991Z Evidence SHA-256 `a2223beea99910879fd4c5ded70604d3d62dcff1f04a3b617d13f28f03a75948` Source locator `/items/6` |  |  |  |  |  |  |
| [DesignArena Image to HTML](https://benchize.com/benchmarks/designarena%3Aimage-to-html.md) | 1,227elo | — | 12 / 35 | — | [designarena · public](https://www.designarena.ai/api/leaderboard#board=image-to-html) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1227 \| elo \| image-to-html \| Higher is better \| ### Configuration arenaType models category website inputModality image sourceModelVersion gemini-3.8-flash variationName public ### Reported statistics averageGenerationTimeMs 122028 battles 6703 losses 3225 standardError Not reported winRate 51.9 wins 3478 ### Source [Original source](https://www.designarena.ai/api/leaderboard#board=image-to-html) Source designarena Original model label Gemini 3.8 Flash Source version public Observed 2026-10-21T16:37:58.959Z Evidence SHA-256 `94f97ebbdd53a8c56f138b620e59b3a9b35c66c3954dcaaa1b2ed4c71518b404` Source locator `/data/12` |  |  |  |  |  |  |
| [DesignArena Data visualization generation](https://benchize.com/benchmarks/designarena%3Adata-viz.md) | 1,269elo | — | 24 / 99 | — | [designarena · public](https://www.designarena.ai/api/leaderboard#board=data-viz) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1269 \| elo \| data-viz \| Higher is better \| ### Configuration arenaType models category dataviz inputModality Not reported sourceModelVersion gemini-3.8-flash variationName public ### Reported statistics averageGenerationTimeMs 96931 battles 524 losses 257 standardError Not reported winRate 51 wins 267 ### Source [Original source](https://www.designarena.ai/api/leaderboard#board=data-viz) Source designarena Original model label Gemini 3.8 Flash Source version public Observed 2026-10-21T16:37:58.264Z Evidence SHA-256 `a895f6f1a25fa8eb01f821c2097ddec4f884de7bd22f1b8a22e82e6679b71849` Source locator `/data/33` |  |  |  |  |  |  |
| [BrokenArXiv 09/2026](https://benchize.com/benchmarks/matharena%3Aarxiv_false--september.md) | 26.73% | 95% CI 18.31–35.15% | 6 / 6 | 0.73usd per problem | [matharena · arxiv_false--september](https://matharena.ai/competition_tables/arxiv_false--september) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 26.73 \| percent \| competition \| Higher is better \| \| `cost_per_problem` \| 0.73 \| usd_per_problem \| competition \| Lower is better \| ### Configuration competition arxiv_false--september predicted false sourceModelName Gemini 3.8 Flash sourceModelVersion gemini/gemini-38-flash ### Reported statistics Uncertainty 95% confidence interval: 18.31–35.15% confidenceInterval95 8.42 problems 53 ### Source [Original source](https://matharena.ai/competition_tables/arxiv_false--september) Source matharena Original model label Gemini 3.8 Flash Source version arxiv_false--september Observed 2026-10-21T16:52:26.084Z Evidence SHA-256 `4431a852d035438c3451a917deda38fbe7f7e38a2da0903d97e836c3b8776de1` Source locator `/table/rows/7` |  |  |  |  |  |  |
| [OpenVibeEval Accessibility](https://benchize.com/benchmarks/openvibeeval%3Aaccessibility.md) | 74 | — | 13 / 26 | — | [openvibeeval](https://openvibeeval.com/leaderboard/) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `adjusted_score` \| 74 \| score \| sample_size_adjusted_accessibility \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration Gemini 3.8 Flash thinkingEnabled Not reported ### Reported statistics runCount 8 ### Source [Original source](https://openvibeeval.com/leaderboard/) Source openvibeeval Original model label Gemini 3.8 Flash Evaluated 2026-09-08 Observed 2026-10-21T16:32:04.885Z Evidence SHA-256 `9b2e7d6060910e3dc7083a6a7f67377601ce78806b2b6b7693cf8d047dd63fbb` Source locator `/rows/18` |  |  |  |  |  |  |

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