# gpt-oss-20bHigh > Source: Benchize (https://benchize.com) — https://benchize.com/models/gpt-oss-20b-high. When citing or reusing this content, credit "Benchize" and link to https://benchize.com/models/gpt-oss-20b-high. OpenAI Canonical URL: https://benchize.com/models/gpt-oss-20b-high Last updated: Thu, 22 Oct 2026 07:32:33 GMT Reproduction and quotation must name Benchize, link to the canonical HTML URL, and retain individual source citations. OpenAI ## Benchmarks | Benchmark | Score | ± | Rank | Cost | Source | | | --- | --- | --- | --- | --- | --- | --- | | [HMMT Feb 2025](https://benchize.com/benchmarks/matharena%3Ahmmt--hmmt_feb_2025.md) | 76.67% | 95% CI 69.1–84.24% | 15 / 35 | 0.01usd per problem | [matharena · hmmt--hmmt_feb_2025](https://matharena.ai/competition_tables/hmmt--hmmt_feb_2025) | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 76.67 \| percent \| competition \| Higher is better \| \| `cost_per_problem` \| 0.011 \| usd_per_problem \| competition \| Lower is better \| ### Configuration competition hmmt--hmmt_feb_2025 predicted false Reasoning effort high sourceModelName GPT OSS 20B (high) sourceModelVersion openai/oss-20b thinkingEnabled true ### Reported statistics Uncertainty 95% confidence interval: 69.1–84.24% confidenceInterval95 7.57 problems 30 ### Source [Original source](https://matharena.ai/competition_tables/hmmt--hmmt_feb_2025) Source matharena Original model label GPT OSS 20B (high) Source version hmmt--hmmt_feb_2025 Observed 2026-10-22T05:30:12.126Z Evidence SHA-256 `8554225c26eb7a3c2649943ee7cf6b404ec1f7441b20fbdcfccbf35b9a864864` Source locator `/table/rows/27` | | | | | | | | [CMIMC 2025](https://benchize.com/benchmarks/matharena%3Acmimc--cmimc_2025.md) | 73.75% | 95% CI 66.93–80.57% | 13 / 22 | 0.01usd per problem | [matharena · cmimc--cmimc_2025](https://matharena.ai/competition_tables/cmimc--cmimc_2025) | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 73.75 \| percent \| competition \| Higher is better \| \| `cost_per_problem` \| 0.010 \| usd_per_problem \| competition \| Lower is better \| ### Configuration competition cmimc--cmimc_2025 predicted false Reasoning effort high sourceModelName GPT OSS 20B (high) sourceModelVersion openai/oss-20b thinkingEnabled true ### Reported statistics Uncertainty 95% confidence interval: 66.93–80.57% confidenceInterval95 6.82 problems 40 ### Source [Original source](https://matharena.ai/competition_tables/cmimc--cmimc_2025) Source matharena Original model label GPT OSS 20B (high) Source version cmimc--cmimc_2025 Observed 2026-10-22T05:30:12.870Z Evidence SHA-256 `d5d48c1f54a744d7ff2244bae3258963afdb1ab36fb20464ea9cf47138a1555e` Source locator `/table/rows/22` | | | | | | | | [BRUMO 2025](https://benchize.com/benchmarks/matharena%3Abrumo--brumo_2025.md) | 86.67% | 95% CI 80.59–92.75% | 16 / 26 | 0.007usd per problem | [matharena · brumo--brumo_2025](https://matharena.ai/competition_tables/brumo--brumo_2025) | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 86.67 \| percent \| competition \| Higher is better \| \| `cost_per_problem` \| 0.007 \| usd_per_problem \| competition \| Lower is better \| ### Configuration competition brumo--brumo_2025 predicted false Reasoning effort high sourceModelName GPT OSS 20B (high) sourceModelVersion openai/oss-20b thinkingEnabled true ### Reported statistics Uncertainty 95% confidence interval: 80.59–92.75% confidenceInterval95 6.08 problems 30 ### Source [Original source](https://matharena.ai/competition_tables/brumo--brumo_2025) Source matharena Original model label GPT OSS 20B (high) Source version brumo--brumo_2025 Observed 2026-10-22T05:30:12.351Z Evidence SHA-256 `d1d1c126b1feff70b2bcc20a7adbf380d94c2aaab0704704cb01593d4e1958ff` Source locator `/table/rows/27` | | | | | | | | [GPQA Diamond](https://benchize.com/benchmarks/gpqa-diamond.md) | 45.96% | SE ±3.55 pp | 115 / 126 | — | [epoch ai · GPQA diamond](https://epoch.ai/data/benchmark_data.zip#file=gpqa_diamond.csv) | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 0.4595959595959596 \| score_fraction \| Best score (across scorers) \| Higher is better \| \| `mean_score` \| 0.4595959595959596 \| score_fraction \| mean_score \| Higher is better \| ### Configuration benchmarkVersion GPQA diamond reasoningBudget Not reported Reasoning effort high sourceModelVersion gpt-oss-20b_high ### Reported statistics Uncertainty Standard error: ±3.55 pp standardError 0.03550702465131341 ### Source [Original source](https://epoch.ai/data/benchmark_data.zip#file=gpqa_diamond.csv) Source epoch ai Original model label gpt-oss-20b Source version GPQA diamond Evaluated 2026-08-06T22:50:56.000Z Observed 2026-10-22T03:00:05.281Z Evidence SHA-256 `9ef0b938bdd8406b029d595983ee8ad3a80bd9d64b82605b91b5ce39ba650c39` Source locator `/rows/110` | | | | | | | | [AIME 2025](https://benchize.com/benchmarks/aime-2025.md) | 89.17% | 95% CI 83.61–94.73% | 12 / 36 | 0.008usd per problem | [matharena · aime--aime_2025](https://matharena.ai/competition_tables/aime--aime_2025) | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 89.17 \| percent \| competition \| Higher is better \| \| `cost_per_problem` \| 0.008 \| usd_per_problem \| competition \| Lower is better \| ### Configuration competition aime--aime_2025 predicted false Reasoning effort high sourceModelName GPT OSS 20B (high) sourceModelVersion openai/oss-20b thinkingEnabled true ### Reported statistics Uncertainty 95% confidence interval: 83.61–94.73% confidenceInterval95 5.56 problems 30 ### Source [Original source](https://matharena.ai/competition_tables/aime--aime_2025) Source matharena Original model label GPT OSS 20B (high) Source version aime--aime_2025 Observed 2026-10-22T05:30:11.892Z Evidence SHA-256 `fab0abcecc4032852ef2cbd2a84d34ffc42202be358b2994bfaba44ae7a0cd9a` Source locator `/table/rows/21` | | | | | | | | [OTIS Mock AIME 2024–2025](https://benchize.com/benchmarks/epoch-ai%3Aotis_mock_aime_2024_2025.md) | 50.83% | SE ±6.71 pp | 93 / 124 | — | [epoch ai · OTIS Mock AIME 2024-2025](https://epoch.ai/data/benchmark_data.zip#file=otis_mock_aime_2024_2025.csv) | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 0.5083333333333333 \| score_fraction \| Best score (across scorers) \| Higher is better \| \| `mean_score` \| 0.5083333333333333 \| score_fraction \| mean_score \| Higher is better \| ### Configuration benchmarkVersion OTIS Mock AIME 2024-2025 reasoningBudget Not reported Reasoning effort high sourceModelVersion gpt-oss-20b_high ### Reported statistics Uncertainty Standard error: ±6.71 pp standardError 0.06711496995394688 ### Source [Original source](https://epoch.ai/data/benchmark_data.zip#file=otis_mock_aime_2024_2025.csv) Source epoch ai Original model label gpt-oss-20b Source version OTIS Mock AIME 2024-2025 Evaluated 2026-08-27T14:13:32.000Z Observed 2026-10-22T03:00:05.281Z Evidence SHA-256 `b7c9da8c0587c82f1e0d814ceb496c24308380b997a16203b6e5d39eb07f4f50` Source locator `/rows/62` | | | | | | | | [SMT 2025](https://benchize.com/benchmarks/matharena%3Asmt--smt_2025.md) | 83.02% | 95% CI 77.97–88.07% | 15 / 26 | 0.007usd per problem | [matharena · smt--smt_2025](https://matharena.ai/competition_tables/smt--smt_2025) | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 83.02 \| percent \| competition \| Higher is better \| \| `cost_per_problem` \| 0.007 \| usd_per_problem \| competition \| Lower is better \| ### Configuration competition smt--smt_2025 predicted false Reasoning effort high sourceModelName GPT OSS 20B (high) sourceModelVersion openai/oss-20b thinkingEnabled true ### Reported statistics Uncertainty 95% confidence interval: 77.97–88.07% confidenceInterval95 5.05 problems 53 ### Source [Original source](https://matharena.ai/competition_tables/smt--smt_2025) Source matharena Original model label GPT OSS 20B (high) Source version smt--smt_2025 Observed 2026-10-22T05:30:12.647Z Evidence SHA-256 `1b4f6f8bafc1688e4557a024513f68befa31aff48fd0623fb4892a61263433b0` Source locator `/table/rows/26` | | | | | | | --- Cite as: gpt-oss-20bHigh — Benchize, https://benchize.com/models/gpt-oss-20b-high (retrieved 2026-10-22)