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 26.3 #74 of 337
Knowledge 6
AA OmniscienceAccuracy AA Omniscience · Accuracy: Measures correct answers on the AA Omniscience factual knowledge test AA 53.4% #14 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 93% #285 of 321
AA Omniscience Index AA Omniscience Index: Scores factual knowledge while penalizing incorrect answers to encourage admitting uncertainty AA 10.1 #42 of 321
IFBench (instructions) IFBench (instructions): Checks whether the model follows precise formatting and content instructions AA 78% #9 of 248
SimpleQA Verifiedv1 SimpleQA Verified v1: Accuracy on short factual questions with verified answers, answered without web access 66.8% #10 of 83
Mystery Game Puzzles Mystery Game Puzzles: Models choose the best move in puzzle positions from an undisclosed game Epoch 26% #28 of 64
Reasoning 5
ARC-AGI-1public evaluation ARC-AGI-1 · public-eval: Models infer transformations from examples and solve unseen grid puzzles; the evaluation split is explicit ARC 88.3% #33 of 61+1 versions & slices semi-private ARC-AGI-1 · semi-private: Models infer transformations from examples and solve unseen grid puzzles; the evaluation split is explicit ARC 84.7% #34 of 64
ARC-AGI-2public evaluation ARC-AGI-2 · public-eval: Models infer transformations from examples and solve unseen grid puzzles; the evaluation split is explicit ARC 34% #34 of 63+1 versions & slices semi-private ARC-AGI-2 · semi-private: Models infer transformations from examples and solve unseen grid puzzles; the evaluation split is explicit ARC 33.6% #37 of 66
Chess Puzzles Chess Puzzles: Models choose the best next move from a novel chess position Epoch 39% #30 of 68
Math 23
MathArenaAIME 2026 MathArena · AIME 2026: Answer accuracy on the named mathematics competition under the recorded effort MathArena 96.7% #6 of 24+18 versions & slices HMMT February 2026 MathArena · HMMT February 2026: Answer accuracy on the named mathematics competition under the recorded effort MathArena 89.4% #13 of 24AIME 2025 MathArena · aime · AIME 2025: Answer accuracy on the named mathematics competition under the recorded effort MathArena 97.5% #2 of 30Apex MathArena · apex · Apex: Answer accuracy on the named mathematics competition under the recorded effort MathArena 15.6% #14 of 34Apex Shortlist MathArena · apex · Apex Shortlist: Answer accuracy on the named mathematics competition under the recorded effort MathArena 68.6% #15 of 30ArXivMath · December 2025 MathArena · arxiv · 12/2025: Answer accuracy on the named mathematics competition under the recorded effort MathArena 43.4% #5 of 13ArXivMath · January 2026 MathArena · arxiv · 01/2026: Answer accuracy on the named mathematics competition under the recorded effort MathArena 58.2% #10 of 18BRUMO 2025 MathArena · brumo · BRUMO 2025: Answer accuracy on the named mathematics competition under the recorded effort MathArena 100% #1 of 24CMIMC 2025 MathArena · cmimc · CMIMC 2025: Answer accuracy on the named mathematics competition under the recorded effort MathArena 90% #7 of 22Project Euler MathArena · euler · Project Euler: Answer accuracy on the named mathematics competition under the recorded effort MathArena 61.7% #6 of 11HMMT February 2025 MathArena · hmmt · HMMT Feb 2025: Answer accuracy on the named mathematics competition under the recorded effort MathArena 97.5% #2 of 30HMMT November 2025 MathArena · hmmt · HMMT Nov 2025: Answer accuracy on the named mathematics competition under the recorded effort MathArena 93.3% #3 of 14Kangaroo 2025 1-2 MathArena · kangaroo · Kangaroo 2025 1-2: Answer accuracy on the named mathematics competition under the recorded effort MathArena 87.5% #4 of 17Kangaroo 2025 11-12 MathArena · kangaroo · Kangaroo 2025 11-12: Answer accuracy on the named mathematics competition under the recorded effort MathArena 95.8% #7 of 17Kangaroo 2025 3-4 MathArena · kangaroo · Kangaroo 2025 3-4: Answer accuracy on the named mathematics competition under the recorded effort MathArena 66.7% #8 of 17Kangaroo 2025 5-6 MathArena · kangaroo · Kangaroo 2025 5-6: Answer accuracy on the named mathematics competition under the recorded effort MathArena 77.5% #7 of 17Kangaroo 2025 7-8 MathArena · kangaroo · Kangaroo 2025 7-8: Answer accuracy on the named mathematics competition under the recorded effort MathArena 89.2% #8 of 16Kangaroo 2025 9-10 MathArena · kangaroo · Kangaroo 2025 9-10: Answer accuracy on the named mathematics competition under the recorded effort MathArena 98.3% #7 of 16SMT 2025 MathArena · smt · SMT 2025: Answer accuracy on the named mathematics competition under the recorded effort MathArena 92.9% #2 of 24
FrontierMath FrontierMath: Models solve expert-written advanced mathematics problems on the named tier and edition Epoch 35.6% #10 of 40
FrontierMath Tier4 FrontierMath Tier 4: Models solve expert-written advanced mathematics problems on the named tier and edition Epoch 4.2% #16 of 33
FrontierMath Tier 4 (v2) FrontierMath Tier 4 (v2): Models solve expert-written advanced mathematics problems on the named tier and edition Epoch 17.1% #42 of 54
FrontierMath Tiers 1–3 (v2) FrontierMath Tiers 1–3 (v2): Models solve expert-written advanced mathematics problems on the named tier and edition Epoch 51.2% #45 of 69
Science 2
Humanity's Last Exam Humanity's Last Exam: Very hard expert questions across many fields, written to stump today’s models AA 36.6% #58 of 337
CritPt (physics) CritPt (physics): Research-level physics problems that need many careful reasoning steps AA 8.6% #66 of 324
Coding 4
WebDev Arena WebDev Arena: Human pairwise preferences in the stated webdev vote pool Arena 1,440 #54 of 97+1 versions & slices full cohort WebDev Arena · snapshot 2026-10-19 · native board settings: Human pairwise preferences in the stated webdev vote pool Arena 1,440 #54 of 97
DeepSWE1 DeepSWE 1: Agents change repository code to resolve software engineering tasks under the stated DeepSWE edition Datacurve 5.1% #17 of 20
Terminal-Bench Hard Terminal-Bench Hard (version unstated): Agents solve difficult software, system-administration and data-processing tasks in a terminal AA 38.6% #35 of 242
Agents 5
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 64% #10 of 16
APEX-Agents APEX-Agents: Agents complete long-horizon professional-services tasks across realistic applications AA 27.7% #15 of 29
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 27.3% #13 of 21
MCP-Atlasfull 1000 MCP-Atlas · full 1000: Tool orchestration across MCP servers on the stated task split Scale 62% #10 of 16
τ³-BankingAA τ³-Banking · AA: Agents retrieve banking knowledge and execute tools to resolve customer-support workflows AA 20.8% #53 of 134
Multimodal 2
MMMU-Pro (vision) MMMU-Pro (vision): College-level questions that require understanding images: charts, diagrams, photos AA 79.9% #20 of 127
Arena Vision Arena Vision · overall: Human pairwise preferences in the stated vision vote pool Arena 1,285 #21 of 88
Writing 16
Arena Textcreative writing Arena Text · creative writing: Human pairwise preferences in the stated text vote pool Arena 1,457 #20 of 195+14 versions & slices overall Arena Text · overall: Human pairwise preferences in the stated text vote pool Arena 1,473 #30 of 196Chinese Arena Text · chinese: Human pairwise preferences in the stated text vote pool Arena 1,511 #28 of 181English Arena Text · english: Human pairwise preferences in the stated text vote pool Arena 1,467 #30 of 188French Arena Text · french: Human pairwise preferences in the stated text vote pool Arena 1,478 #25 of 149German Arena Text · german: Human pairwise preferences in the stated text vote pool Arena 1,497 #7 of 154instruction following Arena Text · instruction following: Human pairwise preferences in the stated text vote pool Arena 1,437 #47 of 188Japanese Arena Text · japanese: Human pairwise preferences in the stated text vote pool Arena 1,489 #7 of 138Korean Arena Text · korean: Human pairwise preferences in the stated text vote pool Arena 1,443 #16 of 144non-English Arena Text · non english: Human pairwise preferences in the stated text vote pool Arena 1,458 #23 of 188Polish Arena Text · polish: Human pairwise preferences in the stated text vote pool Arena 1,494 #7 of 133Russian Arena Text · russian: Human pairwise preferences in the stated text vote pool Arena 1,480 #19 of 186Spanish Arena Text · spanish: Human pairwise preferences in the stated text vote pool Arena 1,469 #20 of 155overall · full cohort Arena Text · overall · snapshot 2026-10-15 · native board settings: Human pairwise preferences in the stated text vote pool Arena 1,466 #26 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 1,457 #17 of 187
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 55.7% #60 of 95
Health & law 5
MLCR-AA MLCR-AA: Models synthesize long medical records; the headline credits concise answers judged both accurate and complete AA 11.1% #39 of 64
Vals LegalBench Vals LegalBench: Legal reasoning and legal-language understanding on LegalBench Vals 86.9% #8 of 89
Vals MedCode Vals MedCode: Medical coding accuracy on patient encounters Vals 55.9% #5 of 78
Vals MedScribe Vals MedScribe: Accuracy of clinical notes produced from consultation transcripts Vals 69.9% #70 of 80
Vals Legal Research Bench Vals Legal Research Bench: Grounded legal research; a task passes only if all criteria pass Vals 18.3% #49 of 69
Documents 3
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
OmniDocBenchv1.6 · full OmniDocBench · v1.6 full: End-to-end document parsing, averaging text recognition, table structure and formula recognition under the stated full-set protocol OmniDocBench 92.6% #13 of 22
Arena Document Arena Document · overall: Human pairwise preferences in the stated document vote pool Arena 1,413 #31 of 33
Legacy · saturated or no longer reported 20
SWE-bench Pro (public) SWE-bench Pro (public): Agents resolve long-horizon professional repository tasks; public and private splits are separate 34.6% #12 of 19
AIME2025 · AA archived AIME 2025 · AA archived: AA evaluates the 30 integer-answer problems from the 2025 AIME mathematics competition AA 97% #2 of 140
LiveCodeBenchAA archived LiveCodeBench · AA archived: AA evaluates solutions to newly collected competitive-programming problems AA 90.8% #1 of 153
GPQA Diamond GPQA Diamond: Graduate-level biology, chemistry and physics questions that can’t be solved by a quick web search 89.8% #36 of 315+1 versions & slices Epoch GPQA Diamond · source setting 2: Models answer graduate-level science multiple-choice questions Epoch 89.4% #37 of 133
τ²-benchairline · 1.0.1 τ-bench official · 2-airline-1.0.1: Customer-service agents converse with simulated users, retrieve domain knowledge and change state through tools Sierra 82.5% #3 of 6+4 versions & slices overall · 1.0.1 τ-bench official · 2-overall-1.0.1: Customer-service agents converse with simulated users, retrieve domain knowledge and change state through tools Sierra 83.5% #4 of 6retail · 1.0.1 τ-bench official · 2-retail-1.0.1: Customer-service agents converse with simulated users, retrieve domain knowledge and change state through tools Sierra 76.8% #4 of 6telecom · 1.0.1 τ-bench official · 2-telecom-1.0.1: Customer-service agents converse with simulated users, retrieve domain knowledge and change state through tools Sierra 91.2% #3 of 6AA τ²-Bench: Tests service agents that use tools while coordinating with a simulated user AA 80.4% #80 of 245
Video-MME2 · without subtitles/audio Video-MME 2 · without subtitles/audio: Grouped non-linear question scoring for video understanding; this is distinct from mean question accuracy Video-MME 32.9% #1 of 18+1 versions & slices 2 · with subtitles/audio Video-MME 2 · with subtitles/audio: Grouped non-linear question scoring for video understanding; this is distinct from mean question accuracy Video-MME 42.5% #1 of 18
OTIS Mock AIME 20242025 OTIS Mock AIME 2024 2025: Models solve the stated OTIS mock AIME mathematics problems Epoch 95.6% #32 of 130
SWE-bench VerifiedEpoch Swe Bench Verified: Agents modify real repositories to resolve the stated Verified issues Epoch 75.4% #10 of 24
SWE-rebenchall · 2025-12-01 to 2026-03-01 SWE-rebench · all · 2025-12-01 to 2026-03-01: Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 53% #1 of 2+5 versions & slices Go · 2025-12-01 to 2026-03-01 SWE-rebench · go · 2025-12-01 to 2026-03-01: Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 2Java · 2025-12-01 to 2026-03-01 SWE-rebench · java · 2025-12-01 to 2026-03-01: Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 2Python · 2025-12-01 to 2026-03-01 SWE-rebench · python · 2025-12-01 to 2026-03-01: Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 53% #1 of 2Rust · 2025-12-01 to 2026-03-01 SWE-rebench · rust · 2025-12-01 to 2026-03-01: Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 2TypeScript · 2025-12-01 to 2026-03-01 SWE-rebench · typescript · 2025-12-01 to 2026-03-01: Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 2
No benchmark matches
CheaperMiMo v2.6 ProthinkingOpen+35%-52%$0.13No PDF input+4
BetterClaude Haiku 5.5xhigh+17%-82%$0.12Remove sampling parameters+5
CheaperMiniMax M3.1 Flashthinking+31%-77%$0.22No PDF input+1
CheaperDeepSeek V4.1 Flash 1017maxOpen+19%-53%$0.26No PDF input+5
CheaperGLM 5.3 FlashmaxOpen+9%-79%$0.28Defaults to max · Self-hosted+4
GPT-6.1 Sollow+30%+256%$0.13Remove unsupported parameters+5
Kimi K3maxOpen+28%+433%$2.00Sampling parameters are fixed+3
Step 5 PreviewthinkingOpen+27%+27%$1.03No PDF input+4
Muse Spark OpenxhighOpen+20%+78%$0.84Reasoning off leaves output unchanged · Self-hosted / llama.cpp+1