Benchmark By Score Rank
Overall 1
AA Intelligence Index4.3.2 AA Intelligence Index 4.3.2: Artificial Analysis combines tests of knowledge, reasoning, coding and tool use into one score AA 31.3 #53 of 337
Knowledge 4
AA OmniscienceAccuracy AA Omniscience · Accuracy: Measures correct answers on the AA Omniscience factual knowledge test AA 49.6% #23 of 321+1 versions & slices Hallucination Rate AA Omniscience · Hallucination Rate: Measures how often attempted answers are incorrect on the AA Omniscience factual knowledge test AA 84.2% #197 of 321
AA Omniscience Index AA Omniscience Index: Scores factual knowledge while penalizing incorrect answers to encourage admitting uncertainty AA 7.2 #46 of 321
IFBench (instructions) IFBench (instructions): Checks whether the model follows precise formatting and content instructions AA 75.9% #16 of 248
Math 2
FrontierMath FrontierMath: Models solve expert-written advanced mathematics problems on the named tier and edition Epoch 39% #7 of 40
FrontierMath Tier4 FrontierMath Tier 4: Models solve expert-written advanced mathematics problems on the named tier and edition Epoch 14.6% #7 of 33
Science 2
Humanity's Last Exam Humanity's Last Exam: Very hard expert questions across many fields, written to stump today’s models AA 40.7% #42 of 337
CritPt (physics) CritPt (physics): Research-level physics problems that need many careful reasoning steps AA 11.3% #54 of 324
Coding 2
SWE-bench Pro (private) SWE-bench Pro (private): Agents resolve long-horizon professional repository tasks; public and private splits are separate Scale 44.7% #3 of 9
Terminal-Bench Hard Terminal-Bench Hard (version unstated): Agents solve difficult software, system-administration and data-processing tasks in a terminal AA 45.5% #19 of 242
Agents 4 Leads on 2
GDPval-AA (work tasks)2.1 GDPval-AA (work tasks) 2.1: Real work tasks from many professions (reports, spreadsheets, analyses), graded against expert work AA 24.9 #79 of 175
AA-Briefcase1.1 · Elo AA-Briefcase 1.1 · Elo: Agents produce professional deliverables; blind pairwise comparisons produce the source Elo rating AA 646 #89 of 140
MCP-Atlas Public MCPAtlas Public: MCPAtlas Public as reported in this launch evaluation; the post is the authority for the recorded task and grading scope 82.4% #1 of 16
MCP-Atlasfull 1000 MCP-Atlas · full 1000: Tool orchestration across MCP servers on the stated task split Scale 82.2% #1 of 16
Multimodal 2
MMMU-Pro (vision) MMMU-Pro (vision): College-level questions that require understanding images: charts, diagrams, photos AA 80.5% #16 of 127
Arena Vision Arena Vision · overall: Human pairwise preferences in the stated vision vote pool Arena 1,306 #9 of 88
Writing 14
Arena Textcreative writing Arena Text · creative writing: Human pairwise preferences in the stated text vote pool Arena 1,465 #15 of 195+13 versions & slices overall Arena Text · overall: Human pairwise preferences in the stated text vote pool Arena 1,489 #13 of 196Chinese Arena Text · chinese: Human pairwise preferences in the stated text vote pool Arena 1,508 #33 of 181English Arena Text · english: Human pairwise preferences in the stated text vote pool Arena 1,480 #16 of 188French Arena Text · french: Human pairwise preferences in the stated text vote pool Arena 1,496 #12 of 149German Arena Text · german: Human pairwise preferences in the stated text vote pool Arena 1,498 #5 of 154instruction following Arena Text · instruction following: Human pairwise preferences in the stated text vote pool Arena 1,442 #42 of 188Korean Arena Text · korean: Human pairwise preferences in the stated text vote pool Arena 1,459 #12 of 144non-English Arena Text · non english: Human pairwise preferences in the stated text vote pool Arena 1,464 #20 of 188Polish Arena Text · polish: Human pairwise preferences in the stated text vote pool Arena 1,474 #13 of 133Russian Arena Text · russian: Human pairwise preferences in the stated text vote pool Arena 1,467 #24 of 186Spanish Arena Text · spanish: Human pairwise preferences in the stated text vote pool Arena 1,469 #19 of 155overall · full cohort Arena Text · overall · snapshot 2026-10-15 · native board settings: Human pairwise preferences in the stated text vote pool Arena 1,474 #18 of 188creative writing · full cohort Arena Text · creative writing · snapshot 2026-10-15 · native board settings: Human pairwise preferences in the stated text vote pool Arena 1,460 #16 of 187
Health & law 3
Vals LegalBench Vals LegalBench: Legal reasoning and legal-language understanding on LegalBench Vals 84.2% #33 of 89
Vals MedCode Vals MedCode: Medical coding accuracy on patient encounters Vals 51.3% #13 of 78
Vals MedScribe Vals MedScribe: Accuracy of clinical notes produced from consultation transcripts Vals 85.9% #23 of 80
Documents 2
AA-LCR (long context)1.1 AA-LCR (long context) 1.1: Questions that need reading and connecting facts across very long documents (around 100K tokens) AA 78% #68 of 275
Arena Document Arena Document · overall: Human pairwise preferences in the stated document vote pool Arena 1,444 #23 of 33
Legacy · saturated or no longer reported 6
SWE-bench Pro (public) SWE-bench Pro (public): Agents resolve long-horizon professional repository tasks; public and private splits are separate 55% #3 of 19
GPQA Diamond GPQA Diamond: Graduate-level biology, chemistry and physics questions that can’t be solved by a quick web search 88.4% #49 of 315+1 versions & slices Epoch GPQA Diamond · source setting 2: Models answer graduate-level science multiple-choice questions Epoch 89.8% #35 of 133
OTIS Mock AIME 20242025 OTIS Mock AIME 2024 2025: Models solve the stated OTIS mock AIME mathematics problems Epoch 88.9% #48 of 130
τ²-benchAA τ²-Bench: Tests service agents that use tools while coordinating with a simulated user AA 91.5% #38 of 245
Terminal-Bench2.1 Terminal-Bench 2.1 · source setting 2: Older, easier version of Terminal-Bench; Still widely reported AA 62.2% #49 of 156
No benchmark matches
