Updated daily at 05:00 UTC

Language

gpt-oss-120b

By OpenAI, in the gpt-oss family.

#85Knowledge & reasoning

Released: 5 Aug 2025
Status: Current
Context: 131.1K
Max output: 131.1K
Input per 1M tokens: —
Output per 1M tokens: —
Open weights: Yes
License: apache-2.0
Parameters: 117B, 5.1B active
Knowledge cutoff: Jun 2024
Reasoning: Yes
Modalities: text → text

Category scores

Out of 100, from independent and third-party results only, by method category-2026-09-29.1.

Category scores
Category	Score	Rank	Coverage
Knowledge & reasoning	16.1	85	3 of 4
Coding & software engineering	41.8	Not ranked	1 of 5
Human preference	43.1	Not ranked	1 of 5

Benchmark results

Every result we hold for this model, with who ran it and where it was published.

Knowledge & reasoning

Knowledge & reasoning results
Benchmark	Metric	Value	Provenance	Setting	Evaluated	Source
Chess Puzzles	Puzzles solved	20.0%	Independent	high	11 Dec 2025	Epoch AI
DTBench (Decision Theory Benchmark)	Accuracy (capability questions)	76.3%	Third party	high	—	https://conceptualreasoning.ai/dtbench via Epoch AI
Humanity's Last Exam (HLE)	Accuracy	19.0%	Self-reported	—	—	GPT-OSS Model Card via Hugging Face Community Evals
LMCA (Language Model Conceptual Argumentation)	Correlation with expert ratings (Pearson r)	0.2214	Third party	high	—	https://conceptualreasoning.ai/lmca via Epoch AI
MMLU-Pro	Accuracy	80.8%	Self-reported	—	—	EvalEval via Hugging Face Community Evals
Mystery Game Puzzles	Puzzles solved	2.0%	Independent	medium	27 Aug 2026	Epoch AI
Mystery Game Puzzles	Puzzles solved	0.0%	Independent	high	27 Aug 2026	Epoch AI
SimpleBench	Score (AVG@5)	22.1%	Third party	—	—	SimpleBench Leaderboard via Epoch AI

Scientific reasoning

Scientific reasoning results
Benchmark	Metric	Value	Provenance	Setting	Evaluated	Source
GPQA Diamond	Accuracy	75.8%	Independent	high	11 Dec 2025	Epoch AI
GPQA Diamond	Accuracy	80.9%	Self-reported	—	—	GPT-OSS Model Card via Hugging Face Community Evals
Surface Evolver Bench	Mean score (partial credit)	25.0%	Third party	—	—	https://yhenon.github.io/surface-evolver-llm-eval/ via Epoch AI

Mathematics

Mathematics results
Benchmark	Metric	Value	Provenance	Setting	Evaluated	Source
OTIS Mock AIME 2024-2025	Accuracy	88.9%	Independent	high	11 Dec 2025	Epoch AI

Coding & software engineering

Coding & software engineering results
Benchmark	Metric	Value	Provenance	Setting	Evaluated	Source
Aider Polyglot	Percent correct	41.8%	Third party	high	—	https://aider.chat/docs/leaderboards/ via Epoch AI
Aider Polyglot	Percent correct	41.8%	Third party	high	—	Aider LLM Leaderboards via Epoch AI
METR Time Horizons	Average task success (Epoch's column)	56.6%	Third party	—	—	METR - Measuring AI Ability to Complete Long Tasks via Epoch AI
Terminal-Bench 2.0	Accuracy (mean over trials)	18.7%	Third party	—	—	Terminal-Bench v2 Leaderboard via Epoch AI
Terminal-Bench 2.0	Accuracy (mean over trials)	18.7%	Third party	unknown	—	https://www.tbench.ai/leaderboard/terminal-bench/2.0 via Epoch AI
Terminal-Bench 2.0	Accuracy (mean over trials)	14.2%	Third party	—	—	Terminal-Bench v2 Leaderboard via Epoch AI
Terminal-Bench 2.0	Accuracy (mean over trials)	14.2%	Third party	unknown	—	https://www.tbench.ai/leaderboard/terminal-bench/2.0 via Epoch AI
WeirdML (v2)	Average accuracy	48.2%	Third party	high	—	WeirdML Leaderboard via Epoch AI
WeirdML (v2)	Average accuracy	48.2%	Third party	high	—	WeirdML Leaderboard via Epoch AI
WeirdML (v2)	Average accuracy	41.9%	Third party	medium	—	https://htihle.github.io/weirdml.html via Epoch AI

Professional knowledge work

Professional knowledge work results
Benchmark	Metric	Value	Provenance	Setting	Evaluated	Source
APEX-Agents	Pass@1 (all rubric criteria met)	4.4%	Source not stated	—	—	Epoch AI via Epoch AI

Long context

Long context results
Benchmark	Metric	Value	Provenance	Setting	Evaluated	Source
Fiction.LiveBench	% correct at 16K tokens (Epoch's column)	44.4%	Third party	high	—	Fiction.live leaderboard via Epoch AI
Fiction.LiveBench	% correct at 16K tokens (Epoch's column)	36.1%	Third party	—	—	Fiction.live leaderboard via Epoch AI

Writing

Writing results
Benchmark	Metric	Value	Provenance	Setting	Evaluated	Source
LLM Creative Story-Writing Benchmark V1 (Lech Mazur)	Mean rubric score (percent of the 10-point scale)	77.1%	Third party	—	—	lechmazur/writing Github repository via Epoch AI

Human preference

Human preference results
Benchmark	Metric	Value	Provenance	Setting	Evaluated	Source
Text Arena (arena.ai)	Arena score (Elo scale)	1352	Independent	—	25 Sep 2026	Arena
Text Arena (arena.ai)	Votes	30021	Independent	—	25 Sep 2026	Arena

Other gpt-oss versions

Sources

gpt-oss-120b: benchmarks, price and specs · supermarket.dev