RankingEffort benchmarks

gpt-oss-20b effort benchmarks

Compare measured Low, Medium, High and other settings. Every number keeps its source.

Model

gpt-oss-20b

45 measured results · 3 reported effort labels · 26 benchmark/harness combinations. Model evidence and sources →

Capability score evidence

gpt-oss-20b · Medium: 4.2 overall comparison score. Meets the overall evidence requirements.

Which sources carry the weight?

Epoch AI · 100.0% direct comparison weight · independent

These shares describe the comparisons entering the fit, not fractions of the final score.

How sensitive is this score?

Remove one evaluator or family and refit the graph against the same overall panel. Values use the matching dataset audited 2026-09-12. These are sensitivity checks, not confidence intervals; ranking eligibility may also change.

Removed evidence	Recomputed score	Change
Artificial Analysis	4.7	+0.5
ARC Prize	5.2	+1.0
Datacurve	3.6	-0.6
LiveCodeBench	4.1	-0.0
Scale	4.6	+0.4
Terminal-Bench	4.2	+0.0
Cognition	4.5	+0.3
LiveBench	4.2	+0.0
SWE-rebench	4.2	+0.0
Epoch AI	Comparison path lost	—
Vals AI	4.1	-0.1
Provider reports · Anthropic	4.2	+0.0
Provider reports · Moonshot AI	4.2	-0.0
Provider reports · Meta	4.2	-0.0
Provider reports · Mistral	4.2	+0.0
All provider reports	4.2	-0.0

Every benchmark family

Removed evidence	Recomputed score	Change
aime	2.6	-1.6
gpqa	5.7	+1.5
arc-agi	5.2	+1.0
deepswe	3.6	-0.6
epoch-game-puzzles	3.7	-0.4
frontiercode	4.5	+0.3
mmmu-pro	4.4	+0.3
simpleqa	4.4	+0.2
hle	4.4	+0.2
tau-bench	4.0	-0.2
mmlu-pro	4.3	+0.1
critpt	4.1	-0.1
livebench-instructions	4.2	+0.1
multi-swe-bench	4.1	-0.1
harvey	4.1	-0.0
swe-rebench	4.2	+0.0
livebench-language	4.2	+0.0
livebench-coding	4.1	-0.0
scicode	4.2	+0.0
swe-atlas-qna	4.1	-0.0
apex-agents	4.2	+0.0
livebench-reasoning	4.1	-0.0
frontiermath	4.1	-0.0
gdp-pdf	4.2	+0.0
vibe-code	4.1	-0.0
arc-agi-3	4.1	-0.0
aa-lcr	4.2	+0.0
gmmlu	4.2	+0.0
livecodebench	4.2	+0.0
vals-finance-agent	4.1	-0.0
vals-legal-research	4.1	-0.0
gdpval	4.1	-0.0
swe-atlas-test-writing	4.2	-0.0
vals-code-migration	4.2	-0.0
omniscience	4.2	-0.0
enigma-eval	4.2	+0.0
livebench-math	4.2	+0.0
livebench-data	4.2	+0.0
analyst-agent	4.2	+0.0
ifbench	4.2	+0.0
math500	4.2	+0.0
itbench	4.2	-0.0
briefcase	4.2	-0.0
vals-excel-modeling	4.2	-0.0
swe-atlas-refactoring	4.2	-0.0
enterprise-ops	4.2	+0.0
automationbench	4.2	-0.0
terminal-bench	4.2	-0.0
swe-bench-pro	4.2	+0.0

What evidence is missing from the fit?

Across all collected settings for this model: 46 observations, 43 contributing, 0 matched without graph weight and 3 excluded. Counts do not establish rank eligibility.

2 · Benchmark family or source protocol has not been reviewed
1 · No reviewed effort for this catalog observation; any separately collected effort measurement is counted separately

Inspect every observation and exclusion →

Capability profile

One configuration, seven capabilities

Comparison score · 0–100

Agpt-oss-20b · Medium1 of 7 capabilities supported

Gaps are unknown, not zero. Hollow points are preliminary.

Agentic
Unknown
Hard reasoning
A5.7Supported
Coding
Unknown
Human pref
Unknown
Knowledge
Unknown
Multimodal
Unknown
Long context
Unknown

Each profile belongs to the named configuration. Capability scores use a shared panel for all axes, distinct from the overall ranking panel; they are not accuracy percentages or direct head-to-head wins.

Compare this configuration →Explore the model evidence profile →

These are published benchmark scores. Supported effort estimates also appear on the Capability leaderboard. Results from different harnesses occupy different rows. Missing cells stay unknown; no scores are copied between effort levels.

Benchmark / source	Low	Medium	High
AA-Briefcase EloArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details AA-Briefcase is an agentic knowledge work benchmark developed by Artificial Analysis. AA-Briefcase Elo is a combined metric that aggregates rubric pass rate, analytical quality Elo and presentation Elo · Higher is better · Evaluation results measured independently by Artificial Analysis Unit: elo. Source collected 2026-09-12.	Not reported	Not reported	0 Elo Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
AIME 2025Artificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	62.33% Reported settings gpt-oss-20b (low); Model source: https://artificialanalysis.ai/models/gpt-oss-20b-low	Not reported	89.33% Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
APEX-Agents-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	Not reported	Not reported	0.74% Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
AA-LCR v1.1Artificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	31% Reported settings gpt-oss-20b (low); Model source: https://artificialanalysis.ai/models/gpt-oss-20b-low	Not reported	34.67% Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
AutomationBench-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of task objectives completed with no guardrail violations · Higher is better · Benchmark developed by Zapier · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	Not reported	Not reported	0.19% Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
CritPtArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark developed by Argonne and UIUC, with contributions from 60+ researchers globally · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	0% Reported settings gpt-oss-20b (low); Model source: https://artificialanalysis.ai/models/gpt-oss-20b-low	Not reported	1.43% Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
GDP.pdf: All-passArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of attempts where every atomic criterion passed · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	Not reported	Not reported	2% Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
GDPval-AA v2 LeaderboardArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Elo rating for performance on real-world work tasks · Anchored to a human baseline of 1,000 · Higher is better · Evaluation results measured independently by Artificial Analysis Unit: elo. Source collected 2026-09-12.	Not reported	Not reported	513.93 Elo Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
Global-MMLU-LiteArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	72.85% Reported settings gpt-oss-20b (low); Model source: https://artificialanalysis.ai/models/gpt-oss-20b-low	Not reported	75.16% Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
GPQA DiamondArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	61.11% Reported settings gpt-oss-20b (low); Model source: https://artificialanalysis.ai/models/gpt-oss-20b-low	Not reported	68.79% Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
Humanity's Last ExamArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	5.28% Reported settings gpt-oss-20b (low); Model source: https://artificialanalysis.ai/models/gpt-oss-20b-low	Not reported	10.98% Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
IFBenchArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	57.82% Reported settings gpt-oss-20b (low); Model source: https://artificialanalysis.ai/models/gpt-oss-20b-low	Not reported	65.1% Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
LiveCodeBenchArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	65.19% Reported settings gpt-oss-20b (low); Model source: https://artificialanalysis.ai/models/gpt-oss-20b-low	Not reported	77.67% Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
MLCR-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of tasks judged accurate, complete, and concise · Expert + compound (hard) sets · Benchmark developed by Wisedocs · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	Not reported	Not reported	0.56% Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
MMLU-ProArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	71.8% Reported settings gpt-oss-20b (low); Model source: https://artificialanalysis.ai/models/gpt-oss-20b-low	Not reported	74.77% Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
AA-Omniscience IndexArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details AA-Omniscience Index (higher is better) measures knowledge reliability and hallucination. It rewards correct answers, penalizes hallucinations, and has no penalty for refusing to answer. Scores range from -100 to 100, where 0 means as many correct as incorrect answers, and negative scores mean more incorrect than correct. · Evaluation results measured independently by Artificial Analysis Unit: index. Source collected 2026-09-12.	-58.55 Reported settings gpt-oss-20b (low); Model source: https://artificialanalysis.ai/models/gpt-oss-20b-low	Not reported	-63.05 Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
AA-Omniscience AccuracyArtificial Analysis · 6000 questions; 1 attempt; correct-answer share; GPT-5.6 Luna medium grader Benchmark details AA-Omniscience Accuracy (higher is better) measures the proportion of correctly answered questions out of all questions, regardless of whether the model chooses to answer · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	15.08% Reported settings gpt-oss-20b (low); Model source: https://artificialanalysis.ai/models/gpt-oss-20b-low	Not reported	16% Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
SciCodeArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	Not reported	Not reported	38.89% Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
𝜏²-Bench TelecomArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	50.29% Reported settings gpt-oss-20b (low); Model source: https://artificialanalysis.ai/models/gpt-oss-20b-low	Not reported	60.23% Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
𝜏³-BankingArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark developed by Sierra Research · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	Not reported	Not reported	7.01% Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
Terminal-Bench HardArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	4.55% Reported settings gpt-oss-20b (low); Model source: https://artificialanalysis.ai/models/gpt-oss-20b-low	Not reported	10.61% Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
Terminal-Bench v2.1Artificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark developed by the Laude Institute and open-source contributors · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	Not reported	Not reported	13.86% Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
Terminal-Bench v4.0Artificial Analysis · mini-SWE-agent v2.4.6; 66 tasks; pass@1 averaged over 3 repeats; 500 steps; task timeout up to 8 hours Benchmark details Benchmark developed by the Laude Institute, Stanford, and open-source contributors · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	Not reported	Not reported	0% Reported settings gpt-oss-20b (high); Model source: https://artificialanalysis.ai/models/gpt-oss-20b
Epoch · Chess Puzzles · v1.1.6Epoch AI · Epoch AI Inspect; task 1.1.6; scorer model_extracted_exact_match Benchmark details Epoch AI internal evaluation. Mean accuracy under the published setting; exact task version and scorer stay separate. Related tasks share a family budget. Data by Epoch AI, CC BY. Unit: percent. Source collected 2026-09-12.	2% Reported settings Epoch AI Inspect; task 1.1.6; scorer model_extracted_exact_match; model=gpt-oss-20b_low; run=QX76VGMs6CYv9BGBKMCCNW; mean accuracy; https://epoch.ai/data/benchmarks.csv	4% Reported settings Epoch AI Inspect; task 1.1.6; scorer model_extracted_exact_match; model=gpt-oss-20b_medium; run=9F6GMXWPSzWo3evzcpAkTd; mean accuracy; https://epoch.ai/data/benchmarks.csv	0% Reported settings Epoch AI Inspect; task 1.1.6; scorer model_extracted_exact_match; model=gpt-oss-20b_high; run=fjgo9DvKNgVAGtitPvgpY7; mean accuracy; https://epoch.ai/data/benchmarks.csv
Epoch · GPQA diamond · v1.0.11Epoch AI · Epoch AI Inspect; task 1.0.11; scorer choice Benchmark details Epoch AI internal evaluation. Mean accuracy under the published setting; exact task version and scorer stay separate. Related tasks share a family budget. Data by Epoch AI, CC BY. Unit: percent. Source collected 2026-09-12.	53.16% Reported settings Epoch AI Inspect; task 1.0.11; scorer choice; model=gpt-oss-20b_low; run=DCJQMqipndGxwMXe8BDGeg; mean accuracy; https://epoch.ai/data/benchmarks.csv	60.8% Reported settings Epoch AI Inspect; task 1.0.11; scorer choice; model=gpt-oss-20b_medium; run=GEwyLqZT93Hp55fdtHrJhm; mean accuracy; https://epoch.ai/data/benchmarks.csv	45.96% Reported settings Epoch AI Inspect; task 1.0.11; scorer choice; model=gpt-oss-20b_high; run=hYATFPeUnMq7Std2nLUVAJ; mean accuracy; https://epoch.ai/data/benchmarks.csv
Epoch · OTIS Mock AIME 2024-2025 · v1.1.6Epoch AI · Epoch AI Inspect; task 1.1.6; scorer model_extracted_exact_match Benchmark details Epoch AI internal evaluation. Mean accuracy under the published setting; exact task version and scorer stay separate. Related tasks share a family budget. Data by Epoch AI, CC BY. Unit: percent. Source collected 2026-09-12.	40.28% Reported settings Epoch AI Inspect; task 1.1.6; scorer model_extracted_exact_match; model=gpt-oss-20b_low; run=cMwETYmC6b4yKupyJfEHv2; mean accuracy; https://epoch.ai/data/benchmarks.csv	65.28% Reported settings Epoch AI Inspect; task 1.1.6; scorer model_extracted_exact_match; model=gpt-oss-20b_medium; run=DB8ja6pEZWGocQbdciyisL; mean accuracy; https://epoch.ai/data/benchmarks.csv	50.83% Reported settings Epoch AI Inspect; task 1.1.6; scorer model_extracted_exact_match; model=gpt-oss-20b_high; run=GMngscbSmqoPbGZWNv2Jvw; mean accuracy; https://epoch.ai/data/benchmarks.csv

What these comparisons mean

Each source reports its own effort labels. Matching labels do not establish equal compute budgets or identical fallback behavior. Open Reported settings to inspect the exact model name, agent and task coverage. Multiple results at one level remain visible rather than selecting the best. Source-reported None or Non-reasoning labels are not treated as evidence that an API supports those settings.

These measurements use their own dated source collection. The leaderboard fits identified configurations jointly; it contains no pooled model ratings. Explicit thinking budgets are separate configurations. Read the methodology.

Download all effort benchmark results · 7,424 results across 240 models · Collected 2026-09-12

gpt-oss-20b effort benchmarks

gpt-oss-20b

Capability score evidence

One configuration, seven capabilities

Agentic

Hard reasoning

Coding

Human pref

Knowledge

Multimodal

Long context