Rankings / OpenAI
GPT OSS 120B
released 2025-08-05
BenchAtlas Index
as of 2026-09-11
high effort
rank #309
10 families · 5 categories · high
base configuration
rank #317
4 families · 3 categories · medium
base configuration
rank #327
9 families · 4 categories · medium
Benchmark evidence
68 results
agentic coding
| τ²-Bench | high effort | 65.8 % | independent | Artificial Analysis |
| τ²-Bench | low effort | 45.0 % | independent | Artificial Analysis |
| τ²-Bench subset=banking | high effort | 12.8 % | independent | Artificial Analysis |
| τ²-Bench subset=banking | low effort | 2.9 % | independent | Artificial Analysis |
coding
| LiveCodeBench v6 implementation=artificial-analysis | high effort | 87.8 % | independent | Artificial Analysis |
| LiveCodeBench v6 implementation=artificial-analysis | low effort | 70.7 % | independent | Artificial Analysis |
| SciCode | low effort | 36.0 % | independent | Artificial Analysis |
| SciCode | high effort | 34.0 % | independent | Artificial Analysis |
external indices
| AA Coding Index | high effort | 30.4 points | independent | Artificial Analysis |
| AA Coding Index | low effort | 21.2 points | independent | Artificial Analysis |
| Intelligence Index v4.1 | high effort | 12.3 points | independent | Artificial Analysis |
| Intelligence Index v4.1 | low effort | 10.2 points | independent | Artificial Analysis |
| AA Math Index | high effort | 93.4 points | independent | Artificial Analysis |
| AA Math Index | low effort | 66.7 points | independent | Artificial Analysis |
| ECI | 140.1 points | independent | Epoch AI Benchmarking Hub | |
| ECI | medium effort | 140.1 points | independent | Epoch AI Benchmarking Hub |
| ECI | high effort | 140.1 points | independent | Epoch AI Benchmarking Hub |
factuality
| MASK contamination=Potential contamination warning: This model was evaluated after the public release of MASK, allowing model builder access to the prompts and solutions. | 92.0 % 91.1–92.9 | independent | Scale Labs |
human preference
| Coding (style control)older version arena=text · category=coding · style_control=true | 1390.8 1383.0–1398.5 | community | LMArena Leaderboard Dataset | |
| Industry Software And It Services (style control)older version arena=text · category=industry_software_and_it_services · style_control=true | 1386.3 1379.9–1392.8 | community | LMArena Leaderboard Dataset | |
| Industry Mathematical (style control)older version arena=text · category=industry_mathematical · style_control=true | 1381.9 1367.0–1396.8 | community | LMArena Leaderboard Dataset | |
| Math (style control)older version arena=text · category=math · style_control=true | 1379.6 1365.8–1393.4 | community | LMArena Leaderboard Dataset | |
| Chinese (style control)older version arena=text · category=chinese · style_control=true | 1377.1 1364.3–1389.9 | community | LMArena Leaderboard Dataset | |
| Hard Prompts English (style control)older version arena=text · category=hard_prompts_english · style_control=true | 1366.9 1359.2–1374.6 | community | LMArena Leaderboard Dataset | |
| Industry Medicine And Healthcare (style control)older version arena=text · category=industry_medicine_and_healthcare · style_control=true | 1365.8 1350.7–1381.0 | community | LMArena Leaderboard Dataset | |
| Spanish (style control)older version arena=text · category=spanish · style_control=true | 1365.4 1344.1–1386.8 | community | LMArena Leaderboard Dataset | |
| Hard Prompts (style control)older version arena=text · category=hard_prompts · style_control=true | 1362.1 1356.4–1367.8 | community | LMArena Leaderboard Dataset | |
| English (style control)older version arena=text · category=english · style_control=true | 1361.7 1355.8–1367.6 | community | LMArena Leaderboard Dataset | |
| Industry Life And Physical And Social Science (style control)older version arena=text · category=industry_life_and_physical_and_social_science · style_control=true | 1360.1 1351.2–1368.9 | community | LMArena Leaderboard Dataset | |
| Expert (style control)older version arena=text · category=expert · style_control=true | 1354.6 1338.3–1370.9 | community | LMArena Leaderboard Dataset | |
| Overall (style control) arena=text · category=overall · style_control=true | 1352.5 1348.1–1356.8 | community | LMArena Leaderboard Dataset | |
| Industry Business And Management And Financial Operations (style control)older version arena=text · category=industry_business_and_management_and_financial_operations · style_control=true | 1351.6 1343.2–1360.0 | community | LMArena Leaderboard Dataset | |
| Industry Legal And Government (style control)older version arena=text · category=industry_legal_and_government · style_control=true | 1341.0 1327.1–1354.8 | community | LMArena Leaderboard Dataset | |
| Non English (style control)older version arena=text · category=non_english · style_control=true | 1337.9 1332.5–1343.2 | community | LMArena Leaderboard Dataset | |
| Russian (style control)older version arena=text · category=russian · style_control=true | 1334.1 1321.6–1346.6 | community | LMArena Leaderboard Dataset | |
| Polish (style control)older version arena=text · category=polish · style_control=true | 1331.9 1317.2–1346.5 | community | LMArena Leaderboard Dataset | |
| German (style control)older version arena=text · category=german · style_control=true | 1329.1 1306.1–1352.1 | community | LMArena Leaderboard Dataset | |
| Japanese (style control)older version arena=text · category=japanese · style_control=true | 1328.0 1302.7–1353.3 | community | LMArena Leaderboard Dataset | |
| Multi Turn (style control)older version arena=text · category=multi_turn · style_control=true | 1327.0 1318.2–1335.8 | community | LMArena Leaderboard Dataset | |
| Longer Query (style control)older version arena=text · category=longer_query · style_control=true | 1324.9 1317.1–1332.7 | community | LMArena Leaderboard Dataset | |
| Instruction Following (style control)older version arena=text · category=instruction_following · style_control=true | 1324.2 1317.1–1331.3 | community | LMArena Leaderboard Dataset | |
| Exclude Ties (style control)older version arena=text · category=exclude_ties · style_control=true | 1313.7 1307.6–1319.9 | community | LMArena Leaderboard Dataset | |
| Industry Writing And Literature And Language (style control)older version arena=text · category=industry_writing_and_literature_and_language · style_control=true | 1309.3 1301.6–1317.0 | community | LMArena Leaderboard Dataset | |
| Industry Entertainment And Sports And Media (style control)older version arena=text · category=industry_entertainment_and_sports_and_media · style_control=true | 1285.1 1276.8–1293.3 | community | LMArena Leaderboard Dataset | |
| Creative Writing (style control)older version arena=text · category=creative_writing · style_control=true | 1277.7 1267.8–1287.7 | community | LMArena Leaderboard Dataset | |
| Korean (style control)older version arena=text · category=korean · style_control=true | 1265.1 1240.6–1289.6 | community | LMArena Leaderboard Dataset |
knowledge science
| GPQA Diamond implementation=vals-ai | 78.5 % | independent | Vals AI | |
| GPQA Diamond implementation=artificial-analysis | high effort | 78.2 % | independent | Artificial Analysis |
| GPQA Diamond implementation=artificial-analysis | low effort | 67.2 % | independent | Artificial Analysis |
| Humanity's Last Exam implementation=artificial-analysis | high effort | 19.6 % | independent | Artificial Analysis |
| Humanity's Last Exam implementation=artificial-analysis | low effort | 5.9 % | independent | Artificial Analysis |
| MMLU-Pro implementation=artificial-analysis | high effort | 80.8 % | independent | Artificial Analysis |
| MMLU-Pro implementation=vals-ai | 79.2 % | independent | Vals AI | |
| MMLU-Pro implementation=artificial-analysis | low effort | 77.5 % | independent | Artificial Analysis |
long context instruction
| AA-LCR | high effort | 52.0 % | independent | Artificial Analysis |
| AA-LCR | low effort | 46.0 % | independent | Artificial Analysis |
| IFBench | high effort | 69.0 % | independent | Artificial Analysis |
| IFBench | low effort | 58.3 % | independent | Artificial Analysis |
| MultiChallenge | 45.3 % 43.2–47.5 | independent | Scale Labs |
professional
| CaseLaw | 48.8 % | independent | Vals AI | |
| CorpFin | 58.2 % | independent | Vals AI | |
| LegalBench | 75.9 % | independent | Vals AI | |
| MedQA | 91.4 % | independent | Vals AI | |
| TaxEval | 71.6 % | independent | Vals AI |
reasoning math
| AIME year=2025 · implementation=artificial-analysis | high effort | 93.4 % | independent | Artificial Analysis |
| AIME implementation=vals-ai | 92.6 % | independent | Vals AI | |
| AIME year=2025 · implementation=artificial-analysis | low effort | 66.7 % | independent | Artificial Analysis |
| MATH-500 implementation=vals-ai | 94.8 % | independent | Vals AI |
Agent + model results
systems, not bare-model scores
| agent + model mini-SWE-agent + GPT OSS 120B | SWE-bench Verified | 26.0 % | community | SWE-bench Leaderboard |
| agent + model mini-SWE-agent + GPT OSS 120B | SWE-bench bash-only | 26.0 % | community | SWE-bench Leaderboard |
| agent + model mini-SWE-agent + GPT OSS 120B | SWE-bench bash-only | 0.1 usd_per_task | community | SWE-bench Leaderboard |
| agent + model mini-SWE-agent + GPT OSS 120B | SWE-bench Verified | 0.1 usd_per_task | community | SWE-bench Leaderboard |
| agent + model Artificial Analysis harness + GPT OSS 120B | Terminal-Bench 2.1 | 26.2 % | independent | Artificial Analysis |
| agent + model Artificial Analysis harness + GPT OSS 120B | Terminal-Bench Hard | 23.5 % | independent | Artificial Analysis |
| agent + model Artificial Analysis harness + GPT OSS 120B | Terminal-Bench 2.1 | 13.9 % | independent | Artificial Analysis |
| agent + model Artificial Analysis harness + GPT OSS 120B | Terminal-Bench Hard | 5.3 % | independent | Artificial Analysis |
These scores measure the whole agent system (scaffold, tools, budgets) — they are never merged into the bare model’s numbers.