Rankings / OpenAI
GPT 4.1 (2025-04-14)
released 2025-04-14
BenchAtlas Index
as of 2026-09-11
high effort
rank #328
10 families · 4 categories · medium
base configuration
rank #388
18 families · 6 categories · high
Benchmark evidence
72 results
agentic coding
| τ²-Bench | 47.1 % | independent | Artificial Analysis |
coding
| LiveCodeBench v6 implementation=artificial-analysis | 45.7 % | independent | Artificial Analysis | |
| SciCode | 38.1 % | independent | Artificial Analysis |
external indices
| Intelligence Index v4.1 | 12.7 points | independent | Artificial Analysis | |
| AA Math Index | 34.7 points | independent | Artificial Analysis | |
| ECI | 136.8 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. | 51.1 % 50.1–52.2 | independent | Scale Labs | |
| SimpleQA Verified | 31.1 % | independent | Epoch AI Benchmarking Hub 2026-08-31 |
human preference
| Coding (style control)older version arena=text · category=coding · style_control=true | 1456.5 1449.9–1463.2 | community | LMArena Leaderboard Dataset | |
| Industry Software And It Services (style control)older version arena=text · category=industry_software_and_it_services · style_control=true | 1442.9 1437.5–1448.3 | community | LMArena Leaderboard Dataset | |
| Hard Prompts English (style control)older version arena=text · category=hard_prompts_english · style_control=true | 1438.0 1431.6–1444.4 | community | LMArena Leaderboard Dataset | |
| Industry Legal And Government (style control)older version arena=text · category=industry_legal_and_government · style_control=true | 1437.0 1426.4–1447.7 | community | LMArena Leaderboard Dataset | |
| Industry Medicine And Healthcare (style control)older version arena=text · category=industry_medicine_and_healthcare · style_control=true | 1432.5 1420.7–1444.3 | community | LMArena Leaderboard Dataset | |
| Hard Prompts (style control)older version arena=text · category=hard_prompts · style_control=true | 1431.5 1426.6–1436.4 | community | LMArena Leaderboard Dataset | |
| Multi Turn (style control)older version arena=text · category=multi_turn · style_control=true | 1430.8 1424.0–1437.8 | community | LMArena Leaderboard Dataset | |
| Polish (style control)older version arena=text · category=polish · style_control=true | 1428.0 1416.4–1439.7 | community | LMArena Leaderboard Dataset | |
| Longer Query (style control)older version arena=text · category=longer_query · style_control=true | 1425.9 1419.5–1432.3 | 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 | 1424.9 1418.0–1431.9 | community | LMArena Leaderboard Dataset | |
| English (style control)older version arena=text · category=english · style_control=true | 1422.6 1417.8–1427.5 | community | LMArena Leaderboard Dataset | |
| Chinese (style control)older version arena=text · category=chinese · style_control=true | 1422.5 1411.8–1433.1 | community | LMArena Leaderboard Dataset | |
| French (style control)older version arena=text · category=french · style_control=true | 1421.3 1398.9–1443.8 | 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 | 1414.7 1407.6–1421.7 | community | LMArena Leaderboard Dataset | |
| Overall (style control) arena=text · category=overall · style_control=true | 1414.4 1410.7–1418.1 | community | LMArena Leaderboard Dataset | |
| Spanish (style control)older version arena=text · category=spanish · style_control=true | 1409.6 1391.2–1427.9 | community | LMArena Leaderboard Dataset | |
| German (style control)older version arena=text · category=german · style_control=true | 1409.5 1392.8–1426.1 | community | LMArena Leaderboard Dataset | |
| Russian (style control)older version arena=text · category=russian · style_control=true | 1409.3 1399.9–1418.8 | community | LMArena Leaderboard Dataset | |
| Expert (style control)older version arena=text · category=expert · style_control=true | 1405.0 1393.2–1416.8 | 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 | 1404.8 1398.6–1411.0 | community | LMArena Leaderboard Dataset | |
| Instruction Following (style control)older version arena=text · category=instruction_following · style_control=true | 1402.9 1397.1–1408.7 | community | LMArena Leaderboard Dataset | |
| Exclude Ties (style control)older version arena=text · category=exclude_ties · style_control=true | 1402.7 1397.6–1407.8 | community | LMArena Leaderboard Dataset | |
| Creative Writing (style control)older version arena=text · category=creative_writing · style_control=true | 1402.5 1394.8–1410.2 | community | LMArena Leaderboard Dataset | |
| Non English (style control)older version arena=text · category=non_english · style_control=true | 1402.0 1397.3–1406.6 | 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 | 1400.7 1393.9–1407.4 | community | LMArena Leaderboard Dataset | |
| Math (style control)older version arena=text · category=math · style_control=true | 1374.2 1363.9–1384.5 | community | LMArena Leaderboard Dataset | |
| Industry Mathematical (style control)older version arena=text · category=industry_mathematical · style_control=true | 1372.5 1361.9–1383.2 | community | LMArena Leaderboard Dataset | |
| Korean (style control)older version arena=text · category=korean · style_control=true | 1370.2 1351.4–1389.1 | community | LMArena Leaderboard Dataset | |
| Japanese (style control)older version arena=text · category=japanese · style_control=true | 1347.2 1328.6–1365.8 | community | LMArena Leaderboard Dataset |
knowledge science
| GPQA Diamond | 66.9 % | independent | Epoch AI Benchmarking Hub 2025-04-14 | |
| GPQA Diamond implementation=artificial-analysis | 66.6 % | independent | Artificial Analysis | |
| GPQA Diamond implementation=vals-ai | high effort | 65.4 % | independent | Vals AI |
| Humanity's Last Exam contamination=Potential contamination warning: This model was evaluated after the public release of HLE, allowing model builder access to the prompts and solutions. · implementation=scale | 5.4 % 4.5–6.3 | independent | Scale Labs | |
| Humanity's Last Exam implementation=artificial-analysis | 4.2 % | independent | Artificial Analysis | |
| MMLU-Pro implementation=artificial-analysis | 80.6 % | independent | Artificial Analysis | |
| MMLU-Pro implementation=vals-ai | high effort | 80.5 % | independent | Vals AI |
long context instruction
| AA-LCR | 68.3 % | independent | Artificial Analysis | |
| IFBench | 43.0 % | independent | Artificial Analysis | |
| MultiChallenge | 39.4 % 38.5–40.4 | independent | Scale Labs |
multimodal
| MMMU implementation=vals-ai | high effort | 72.4 % | independent | Vals AI |
| VISTA | 45.3 % 44.4–46.3 | independent | Scale Labs |
professional
| CaseLaw | high effort | 69.9 % | independent | Vals AI |
| CorpFin | high effort | 63.1 % | independent | Vals AI |
| LegalBench | high effort | 83.1 % | independent | Vals AI |
| MedQA | high effort | 91.2 % | independent | Vals AI |
| TaxEval | high effort | 75.1 % | independent | Vals AI |
reasoning math
| AIME implementation=artificial-analysis | 43.7 % | independent | Artificial Analysis | |
| AIME implementation=vals-ai | high effort | 39.6 % | independent | Vals AI |
| AIME year=2025 · implementation=artificial-analysis | 34.7 % | independent | Artificial Analysis | |
| ARC-AGI-1older version split=public_eval · model_type=Base LLM | 11.8 % | independent | ARC Prize Leaderboard | |
| ARC-AGI-1older version split=semi_private · model_type=Base LLM | 5.5 % | independent | ARC Prize Leaderboard | |
| ARC-AGI-2 split=semi_private · model_type=Base LLM | 0.4 % | independent | ARC Prize Leaderboard | |
| ARC-AGI-2 split=semi_private · model_type=Base LLM | 0.1 usd_per_task | independent | ARC Prize Leaderboard | |
| ARC-AGI-2 split=public_eval · model_type=Base LLM | 0.1 usd_per_task | independent | ARC Prize Leaderboard | |
| ARC-AGI-1older version split=semi_private · model_type=Base LLM | 0.0 usd_per_task | independent | ARC Prize Leaderboard | |
| ARC-AGI-1older version split=public_eval · model_type=Base LLM | 0.0 usd_per_task | independent | ARC Prize Leaderboard | |
| ARC-AGI-2 split=public_eval · model_type=Base LLM | 0.0 % | independent | ARC Prize Leaderboard | |
| EnigmaEval contamination=Potential contamination warning: This model was evaluated after the public release of EnigmaEval, allowing model builder access to the prompts and solutions. | 2.2 % 1.3–3.0 | independent | Scale Labs | |
| FrontierMath | 5.5 % | independent | Epoch AI Benchmarking Hub 2025-04-14 | |
| FrontierMath Tier 4older version | 0.0 % | independent | Epoch AI Benchmarking Hub 2025-07-01 | |
| MATH-500 implementation=artificial-analysis | 91.3 % | independent | Artificial Analysis | |
| MATH-500 implementation=vals-ai | high effort | 87.2 % | independent | Vals AI |
| MATH Level 5 | 83.0 % | independent | Epoch AI Benchmarking Hub 2025-04-14 | |
| OTIS Mock AIME 2024–2025 | 38.3 % | independent | Epoch AI Benchmarking Hub 2025-04-14 |
Agent + model results
systems, not bare-model scores
| agent + model Epoch Inspect harness + GPT 4.1 (2025-04-14) | SWE-bench Verified | 48.5 % | independent | Epoch AI Benchmarking Hub |
| agent + model GUIRepair + GPT 4.1 (2025-04-14) | SWE-bench Multimodal | 31.1 % | community | SWE-bench Leaderboard |
| agent + model camel + GPT 4.1 (2025-04-14) | Terminal-Bench 1.0 | 35.0 % | unverified | Terminal-Bench Leaderboard |
| agent + model terminus-1 + GPT 4.1 (2025-04-14) | Terminal-Bench 1.0 | 30.3 % | community | Terminal-Bench Leaderboard |
| agent + model Artificial Analysis harness + GPT 4.1 (2025-04-14) | Terminal-Bench Hard | 13.6 % | independent | Artificial Analysis |
| agent + model Codex CLI + GPT 4.1 (2025-04-14) | Terminal-Bench 1.0 | 8.3 % | community | Terminal-Bench Leaderboard |
These scores measure the whole agent system (scaffold, tools, budgets) — they are never merged into the bare model’s numbers.