Benchmark By Score Rank
Overall 1
AA Intelligence Index4.3.2 Artificial Analysis combines tests of knowledge, reasoning, coding and tool use into one score AA 30.7 #55 of 339
Knowledge 7
AA OmniscienceAccuracy Measures correct answers on the AA Omniscience factual knowledge test AA 54.9% #9 of 323+1 versions & slices Hallucination Rate Measures how often attempted answers are incorrect on the AA Omniscience factual knowledge test AA 88% #231 of 323
AA Omniscience Index Scores factual knowledge while penalizing incorrect answers to encourage admitting uncertainty AA 15.1 #32 of 323
IFBench (instructions) Checks whether the model follows precise formatting and content instructions AA 64.4% #72 of 248
EQ-Bench3 Judge-rated responses to social and emotionally nuanced scenarios, scored through the source pairwise ranking EQ-Bench 1,554 #9 of 53
SimpleQA Verifiedv1 Accuracy on short factual questions with verified answers, answered without web access 63% #14 of 84
Mystery Game Puzzles Models choose the best move in puzzle positions from an undisclosed game Epoch 28% #27 of 65
Reasoning 6
ARC-AGI-1public evaluation Models infer transformations from examples and solve unseen grid puzzles; the evaluation split is explicit ARC 82.3% #37 of 64+1 versions & slices semi-private Models infer transformations from examples and solve unseen grid puzzles; the evaluation split is explicit ARC 76.2% #40 of 67
ARC-AGI-2public evaluation Models infer transformations from examples and solve unseen grid puzzles; the evaluation split is explicit ARC 34.3% #36 of 66+1 versions & slices semi-private Models infer transformations from examples and solve unseen grid puzzles; the evaluation split is explicit ARC 33.3% #40 of 69
ARC-AGI-3semi-private Measures interactive game solving through relative human action efficiency ARC 0.4% #14 of 19
Chess Puzzles Models choose the best next move from a novel chess position Epoch 30% #42 of 69
Math 27 Leads on 1
AA Analyst Agent Data-analyst tasks: the model explores data with tools and reports findings AA 50% #8 of 43
MathArenaAIME 2026 Answer accuracy on the named mathematics competition under the recorded effort MathArena 100% #1 of 24+22 versions & slices HMMT February 2026 Answer accuracy on the named mathematics competition under the recorded effort MathArena 98.5% #1 of 24Apex Answer accuracy on the named mathematics competition under the recorded effort MathArena 80.2% #2 of 34Apex Shortlist Answer accuracy on the named mathematics competition under the recorded effort MathArena 98.4% #1 of 30ArXivMath · April 2026 Answer accuracy on the named mathematics competition under the recorded effort MathArena 68.8% #4 of 17ArXivMath · February 2026 Answer accuracy on the named mathematics competition under the recorded effort MathArena 73.4% #2 of 20ArXivMath · January 2026 Answer accuracy on the named mathematics competition under the recorded effort MathArena 73.9% #2 of 18ArXivMath · June 2026 Answer accuracy on the named mathematics competition under the recorded effort MathArena 83.6% #7 of 24ArXivMath · March 2026 Answer accuracy on the named mathematics competition under the recorded effort MathArena 77.5% #1 of 14ArXivMath · May 2026 Answer accuracy on the named mathematics competition under the recorded effort MathArena 77.5% #7 of 19BrokenArXiv · April 2026 Answer accuracy on the named mathematics competition under the recorded effort MathArena 72.1% #2 of 15BrokenArXiv · February 2026 Answer accuracy on the named mathematics competition under the recorded effort MathArena 68.2% #1 of 15BrokenArXiv · June 2026 Answer accuracy on the named mathematics competition under the recorded effort MathArena 69.4% #7 of 24BrokenArXiv · March 2026 Answer accuracy on the named mathematics competition under the recorded effort MathArena 73.7% #1 of 13BrokenArXiv · May 2026 Answer accuracy on the named mathematics competition under the recorded effort MathArena 50% #6 of 18ArXivLean · March 2026 Answer accuracy on the named mathematics competition under the recorded effort MathArena 17.1% #1 of 5Kangaroo 2025 1-2 Answer accuracy on the named mathematics competition under the recorded effort MathArena 95.8% #1 of 17Kangaroo 2025 11-12 Answer accuracy on the named mathematics competition under the recorded effort MathArena 98.3% #2 of 17Kangaroo 2025 3-4 Answer accuracy on the named mathematics competition under the recorded effort MathArena 89.6% #1 of 17Kangaroo 2025 5-6 Answer accuracy on the named mathematics competition under the recorded effort MathArena 90% #1 of 17Kangaroo 2025 7-8 Answer accuracy on the named mathematics competition under the recorded effort MathArena 95.8% #3 of 16Kangaroo 2025 9-10 Answer accuracy on the named mathematics competition under the recorded effort MathArena 100% #1 of 16USAMO 2026 Answer accuracy on the named mathematics competition under the recorded effort MathArena 98.2% #1 of 8
FrontierMath Erdos Models solve expert-written advanced mathematics problems on the named tier and edition Epoch 0% #6 of 13
FrontierMath Tier 4 (v2) Models solve expert-written advanced mathematics problems on the named tier and edition Epoch 72.5% #12 of 55
FrontierMath Tiers 1–3 (v2) Models solve expert-written advanced mathematics problems on the named tier and edition Epoch 85.3% #13 of 70
Science 3
Humanity's Last Exam Very hard expert questions across many fields, written to stump today’s models AA 32.7% #74 of 339
CritPt (physics) Research-level physics problems that need many careful reasoning steps AA 8% #70 of 326
SciCode Writing code for real scientific computing problems (physics, biology, math) AA 56.1% #25 of 142
Coding 7
Terminal-Bench4.0 Agents use a terminal to complete executable coding, system administration and research tasks in the 4.0 suite 14.6% #33 of 134+1 versions & slices 2.1 Older, easier version of Terminal-Bench; Still widely reported AA 65.5% #42 of 156
DeepSWE1.1 Agents change repository code to resolve software engineering tasks under the stated DeepSWE edition 27% #33 of 44+1 versions & slices 1 Agents change repository code to resolve software engineering tasks under the stated DeepSWE edition Datacurve 70% #1 of 20
ITBench SRE IT incident tasks: diagnose and fix broken systems like a site-reliability engineer AA 45.8% #12 of 37
MirrorCode Agents reimplement software projects by observing their behavior Epoch 10% #9 of 10
Terminal-Bench Hard Agents solve difficult software, system-administration and data-processing tasks in a terminal AA 52.3% #10 of 242
Agents 12
AutomationBench-AA Agents orchestrate SaaS workflows through APIs; objectives receive credit only when the task has no guardrail violation AA 47.3% #41 of 135
AA-Briefcase1.1 · Elo Agents produce professional deliverables; blind pairwise comparisons produce the source Elo rating AA 1,136 #47 of 142
APEX-Agents Agents complete long-horizon professional-services tasks across realistic applications AA 37.7% #5 of 29
EnterpriseOps-Gym-AA Agents complete stateful enterprise workflows in eight domains; final database state determines success AA 46.6% #9 of 43
GDPval-AA (work tasks)2.1 Real work tasks from many professions (reports, spreadsheets, analyses), graded against expert work AA 27.2 #131 of 177
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 79.4% #4 of 16
Toolathlon-Verified Toolathlon-Verified as reported in this launch evaluation; the post is the authority for the recorded task and grading scope 73.5% #6 of 6
τ³-benchbanking · 1.0.1 τ-bench official · 3-banking-1.0.1: Customer-service agents converse with simulated users, retrieve domain knowledge and change state through tools Sierra 44.6% #4 of 21
EBR Bench Agents learn to play Earthborne Rangers across repeated playthroughs Epoch 34.3% #11 of 24
Furniture Assembly Models inspect furniture assembly photos, identify incorrect builds and locate the mistaken step Epoch 44.2% #13 of 31
MCP-Atlasfull 1000 Tool orchestration across MCP servers on the stated task split Scale 75.3% #5 of 16
τ³-BankingAA Agents retrieve banking knowledge and execute tools to resolve customer-support workflows AA 24.9% #42 of 134
Multimodal 3
MMMU-Pro (vision) College-level questions that require understanding images: charts, diagrams, photos AA 79% #26 of 128
Arena Vision Human pairwise preferences in the stated vision vote pool Arena 1,297 #13 of 88
Chartography Expert chart reasoning over professional scientific, medical, financial and engineering charts Surge 31% #13 of 37
Writing 20
Arena Textcreative writing · October 22, 2026 Human pairwise preferences in the stated text vote pool Arena 1,455 #20 of 197+16 versions & slices creative writing Human pairwise preferences in the stated text vote pool Arena 1,451 #25 of 195overall Human pairwise preferences in the stated text vote pool Arena 1,481 #19 of 196Chinese Human pairwise preferences in the stated text vote pool Arena 1,534 #10 of 198English Human pairwise preferences in the stated text vote pool Arena 1,471 #27 of 205French Human pairwise preferences in the stated text vote pool Arena 1,485 #24 of 164German Human pairwise preferences in the stated text vote pool Arena 1,480 #15 of 167instruction following Human pairwise preferences in the stated text vote pool Arena 1,478 #17 of 205Japanese Human pairwise preferences in the stated text vote pool Arena 1,499 #6 of 152Korean Human pairwise preferences in the stated text vote pool Arena 1,460 #9 of 156non-English Human pairwise preferences in the stated text vote pool Arena 1,467 #17 of 205Polish Human pairwise preferences in the stated text vote pool Arena 1,466 #19 of 142Russian Human pairwise preferences in the stated text vote pool Arena 1,473 #21 of 203Spanish Human pairwise preferences in the stated text vote pool Arena 1,468 #23 of 168overall · October 15, 2026 Human pairwise preferences in the stated text vote pool Arena 1,471 #20 of 188creative writing · October 15, 2026 Human pairwise preferences in the stated text vote pool Arena 1,455 #20 of 187overall · October 22, 2026 Human pairwise preferences in the stated text vote pool Arena 1,472 #22 of 198
EQ-Bench Longform Creative Writing1.11 Longform Creative Writing 1.11: Models plan, revise and write a story across eight turns, graded with a writing rubric EQ-Bench 78.2% #21 of 96
EQ-Bench Creative Writingv3 Creative Writing v3 · Elo: Judge-rated creative writing under the v3 prompt set and pairwise ranking EQ-Bench 1,845 #18 of 94
Lech Mazur Short-Story Judges compare stories written to ten required elements; the comparison score can be negative Lech 2.4 #9 of 45
Health & law 5
MLCR-AA Models synthesize long medical records; the headline credits concise answers judged both accurate and complete AA 11.1% #51 of 103
Vals LegalBench Legal reasoning and legal-language understanding on LegalBench Vals 86.5% #10 of 89
Vals MedCode Medical coding accuracy on patient encounters Vals 49.1% #23 of 78
Vals MedScribe Accuracy of clinical notes produced from consultation transcripts Vals 86.9% #19 of 80
Vals Legal Research Bench Grounded legal research; a task passes only if all criteria pass Vals 40.4% #25 of 73
Documents 3
AA-LCR (long context)1.1 Questions that need reading and connecting facts across very long documents (around 100K tokens) AA 81% #35 of 317
GDP.pdfAA all-pass Models reason over professional PDF documents; an attempt passes only when every atomic criterion passes 22.6% #20 of 135
Arena Document Human pairwise preferences in the stated document vote pool Arena 1,486 #6 of 33
Legacy · saturated or no longer reported 4
GPQA Diamond Graduate-level biology, chemistry and physics questions that can’t be solved by a quick web search 91% #29 of 315+1 versions & slices Epoch Models answer graduate-level science multiple-choice questions Epoch 90.7% #29 of 134
OTIS Mock AIME 20242025 Models solve the stated OTIS mock AIME mathematics problems Epoch 84.4% #63 of 131
τ²-benchAA Tests service agents that use tools while coordinating with a simulated user AA 83.9% #67 of 245
No benchmark matches
CheaperClaude Opus 5.5max+21%-29%$5.98Context: 1.05M → 1M+6
OpenDeepSeek V4 Pro 0813max-10%-82%$0.67Context: 1.05M → 1M+6
MiniMax M3.1 Flashthinking+2%-98%$0.22Context: 1.05M → 1M+2
Kimi K3maxOpen-13%-47%$2.00Context: 1.05M → 1M+4
Step 5 PreviewthinkingOpen-24%-87%$1.03Context: 1.05M → 1M+3