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 23.5 #93 of 337
Knowledge 6
AA OmniscienceAccuracy Measures correct answers on the AA Omniscience factual knowledge test AA 35.2% #58 of 321+1 versions & slices Hallucination Rate Measures how often attempted answers are incorrect on the AA Omniscience factual knowledge test AA 65.7% #119 of 321
AA Omniscience Index Scores factual knowledge while penalizing incorrect answers to encourage admitting uncertainty AA -7.3 #86 of 321
IFBench (instructions) Checks whether the model follows precise formatting and content instructions AA 70.2% #50 of 248
EQ-Bench3 Judge-rated responses to social and emotionally nuanced scenarios, scored through the source pairwise ranking EQ-Bench 1,545 #17 of 50
SimpleQA Verifiedv1 Accuracy on short factual questions with verified answers, answered without web access 34.3% #49 of 83
Reasoning 4
ARC-AGI-1public evaluation Models infer transformations from examples and solve unseen grid puzzles; the evaluation split is explicit ARC 73.1% #41 of 61+1 versions & slices semi-private Models infer transformations from examples and solve unseen grid puzzles; the evaluation split is explicit ARC 65.3% #42 of 64
ARC-AGI-2public evaluation Models infer transformations from examples and solve unseen grid puzzles; the evaluation split is explicit ARC 12.1% #41 of 63+1 versions & slices semi-private Models infer transformations from examples and solve unseen grid puzzles; the evaluation split is explicit ARC 11.8% #44 of 66
Math 20
MathArenaAIME 2026 Answer accuracy on the named mathematics competition under the recorded effort MathArena 95.8% #12 of 24+19 versions & slices HMMT February 2026 Answer accuracy on the named mathematics competition under the recorded effort MathArena 87.1% #18 of 24AIME 2025 Answer accuracy on the named mathematics competition under the recorded effort MathArena 95.8% #4 of 30Apex Answer accuracy on the named mathematics competition under the recorded effort MathArena 8.9% #20 of 34Apex Shortlist Answer accuracy on the named mathematics competition under the recorded effort MathArena 57.5% #20 of 30ArXivMath · December 2025 Answer accuracy on the named mathematics competition under the recorded effort MathArena 41.9% #6 of 13ArXivMath · January 2026 Answer accuracy on the named mathematics competition under the recorded effort MathArena 62.5% #9 of 18BrokenArXiv · February 2026 Answer accuracy on the named mathematics competition under the recorded effort MathArena 2.8% #15 of 15BRUMO 2025 Answer accuracy on the named mathematics competition under the recorded effort MathArena 98.3% #4 of 24CMIMC 2025 Answer accuracy on the named mathematics competition under the recorded effort MathArena 90.6% #5 of 22Project Euler Answer accuracy on the named mathematics competition under the recorded effort MathArena 59.9% #8 of 11HMMT February 2025 Answer accuracy on the named mathematics competition under the recorded effort MathArena 93.3% #5 of 30HMMT November 2025 Answer accuracy on the named mathematics competition under the recorded effort MathArena 89.2% #7 of 14Kangaroo 2025 1-2 Answer accuracy on the named mathematics competition under the recorded effort MathArena 76% #9 of 17Kangaroo 2025 11-12 Answer accuracy on the named mathematics competition under the recorded effort MathArena 90% #12 of 17Kangaroo 2025 3-4 Answer accuracy on the named mathematics competition under the recorded effort MathArena 65.6% #10 of 17Kangaroo 2025 5-6 Answer accuracy on the named mathematics competition under the recorded effort MathArena 67.5% #12 of 17Kangaroo 2025 7-8 Answer accuracy on the named mathematics competition under the recorded effort MathArena 88.3% #10 of 16Kangaroo 2025 9-10 Answer accuracy on the named mathematics competition under the recorded effort MathArena 95.8% #9 of 16SMT 2025 Answer accuracy on the named mathematics competition under the recorded effort MathArena 90.6% #7 of 24
Science 2
Humanity's Last Exam Very hard expert questions across many fields, written to stump today’s models AA 30.7% #76 of 337
CritPt (physics) Research-level physics problems that need many careful reasoning steps AA 3.1% #92 of 324
Coding 3
WebDev Arena Human pairwise preferences in the stated webdev vote pool Arena 1,437 #56 of 97+1 versions & slices full cohort Human pairwise preferences in the stated webdev vote pool Arena 1,437 #57 of 97
Terminal-Bench Hard Agents solve difficult software, system-administration and data-processing tasks in a terminal AA 34.8% #44 of 242
Agents 3
GDPval-AA (work tasks)2.1 Real work tasks from many professions (reports, spreadsheets, analyses), graded against expert work AA 17 #96 of 175
APEX-Agents Agents complete long-horizon professional-services tasks across realistic applications AA 11.5% #24 of 29
τ³-BankingAA Agents retrieve banking knowledge and execute tools to resolve customer-support workflows AA 14.2% #73 of 134
Multimodal 3
MMMU-Pro (vision) College-level questions that require understanding images: charts, diagrams, photos AA 75.4% #47 of 127
Arena Vision Human pairwise preferences in the stated vision vote pool Arena 1,269 #33 of 88
Chartography Expert chart reasoning over professional scientific, medical, financial and engineering charts Surge 12.6% #33 of 37
Writing 4
Arena Textcreative writing Human pairwise preferences in the stated text vote pool Arena 1,423 #51 of 195+1 versions & slices overall Human pairwise preferences in the stated text vote pool Arena 1,450 #52 of 196
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 74.9% #27 of 95
EQ-Bench Creative Writingv3 Creative Writing v3 · Elo: Judge-rated creative writing under the v3 prompt set and pairwise ranking EQ-Bench 1,579 #45 of 91
Health & law 3
Vals MedCode Medical coding accuracy on patient encounters Vals 39.3% #53 of 78
Vals MedScribe Accuracy of clinical notes produced from consultation transcripts Vals 76.4% #58 of 80
Vals Legal Research Bench Grounded legal research; a task passes only if all criteria pass Vals 15.9% #51 of 69
Documents 2
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 End-to-end document parsing, averaging text recognition, table structure and formula recognition under the stated full-set protocol OmniDocBench 84.5% #19 of 22
Legacy · saturated or no longer reported 14
GPQA Diamond Graduate-level biology, chemistry and physics questions that can’t be solved by a quick web search 87.9% #51 of 315
Video-MME2 · without subtitles/audio Grouped non-linear question scoring for video understanding; this is distinct from mean question accuracy Video-MME 27.3% #3 of 18+3 versions & slices 2 · with subtitles/audio Grouped non-linear question scoring for video understanding; this is distinct from mean question accuracy Video-MME 36.1% #4 of 182 · with subtitles/audio · 512 frames Grouped non-linear question scoring for video understanding; this is distinct from mean question accuracy Video-MME 40.7% #1 of 12 · without subtitles/audio · 512 frames Grouped non-linear question scoring for video understanding; this is distinct from mean question accuracy Video-MME 33% #1 of 2
SWE-bench VerifiedEpoch Swe Bench Verified: Agents modify real repositories to resolve the stated Verified issues Epoch 73.8% #14 of 24
SWE-rebenchall · 2026-01-01 to 2026-03-01 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 49% #2 of 5+5 versions & slices Go · 2026-01-01 to 2026-03-01 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 5Java · 2026-01-01 to 2026-03-01 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 5Python · 2026-01-01 to 2026-03-01 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 49% #2 of 5Rust · 2026-01-01 to 2026-03-01 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 5TypeScript · 2026-01-01 to 2026-03-01 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 5
τ²-benchAA Tests service agents that use tools while coordinating with a simulated user AA 95.9% #12 of 245
Terminal-Bench2.1 Older, easier version of Terminal-Bench; Still widely reported AA 45.7% #74 of 156
No benchmark matches
CheaperGemini 3 Flash Previewunstated+2%-6%$0.50 / $3
BetterGLM 5.2maxOpen+7%+79%$1.47No image input
CheaperGPT-5 Mini (2025-08-07)medium-12%-43%$0.25 / $2
BetterClaude Opus 4.7max+13%+733%$5 / $25
BetterDeepSeek V4 Pro 0813maxOpen+5%+65%$1.32 / $3.96Memory (4-bit): 616 GB → 990 GB+6
Qwen3.7 Maxmax+5%+84%$1.48 / $4.43No image input