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 33.7 #46 of 337
Knowledge 4
AA OmniscienceAccuracy AA Omniscience · Accuracy: Measures correct answers on the AA Omniscience factual knowledge test AA 52.1% #16 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 50% #88 of 321
AA Omniscience Index AA Omniscience Index: Scores factual knowledge while penalizing incorrect answers to encourage admitting uncertainty AA 28.1 #15 of 321
SimpleQA Verifiedv1 SimpleQA Verified v1: Accuracy on short factual questions with verified answers, answered without web access 57.8% #17 of 83
Reasoning 1
Chess Puzzles Chess Puzzles: Models choose the best next move from a novel chess position Epoch 34% #39 of 68
Math 2
MathArenaArXivMath · June 2026 MathArena · arxiv · 06/2026: Answer accuracy on the named mathematics competition under the recorded effort MathArena 56.9% #19 of 24+1 versions & slices BrokenArXiv · June 2026 MathArena · arxiv_false · 06/2026: Answer accuracy on the named mathematics competition under the recorded effort MathArena 50.6% #12 of 24
Science 3
Humanity's Last Exam Humanity's Last Exam: Very hard expert questions across many fields, written to stump today’s models AA 46.2% #21 of 337
CritPt (physics) CritPt (physics): Research-level physics problems that need many careful reasoning steps AA 15.1% #45 of 324
SciCode SciCode: Writing code for real scientific computing problems (physics, biology, math) AA 58.8% #13 of 140
Coding 6 Leads on 1
Terminal-Bench4.0 Terminal-Bench 4.0 · source setting 2: Agents use a terminal to complete executable coding, system administration and research tasks in the 4.0 suite 6.1% #43 of 132+1 versions & slices 2.1 Terminal-Bench 2.1 · source setting 2: Older, easier version of Terminal-Bench; Still widely reported AA 77.9% #26 of 156
DeepSWE1.1 DeepSWE 1.1: Agents change repository code to resolve software engineering tasks under the stated DeepSWE edition 53.3% #24 of 44
WebDev Arena WebDev Arena: Human pairwise preferences in the stated webdev vote pool Arena 1,542 #37 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 1,542 #37 of 97
SWE-bench Pro (private) SWE-bench Pro (private): Agents resolve long-horizon professional repository tasks; public and private splits are separate Scale 51.5% #1 of 9
Agents 5
AutomationBench-AA AutomationBench-AA: Agents orchestrate SaaS workflows through APIs; objectives receive credit only when the task has no guardrail violation AA 38.8% #43 of 133
AA-Briefcase1.1 · Elo AA-Briefcase 1.1 · Elo: Agents produce professional deliverables; blind pairwise comparisons produce the source Elo rating AA 851 #72 of 140
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 35.6 #51 of 175
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 40.5% #5 of 21
τ³-BankingAA τ³-Banking · AA: Agents retrieve banking knowledge and execute tools to resolve customer-support workflows AA 31.8% #33 of 134
Multimodal 2
Chartography Chartography: Expert chart reasoning over professional scientific, medical, financial and engineering charts Surge 24.4% #20 of 37
Arena Vision Arena Vision · overall: Human pairwise preferences in the stated vision vote pool Arena 1,293 #17 of 88
Writing 18
Arena Textcreative writing Arena Text · creative writing: Human pairwise preferences in the stated text vote pool Arena 1,444 #33 of 195+14 versions & slices overall Arena Text · overall: Human pairwise preferences in the stated text vote pool Arena 1,490 #11 of 196Chinese Arena Text · chinese: Human pairwise preferences in the stated text vote pool Arena 1,516 #24 of 181English Arena Text · english: Human pairwise preferences in the stated text vote pool Arena 1,478 #19 of 188French Arena Text · french: Human pairwise preferences in the stated text vote pool Arena 1,493 #14 of 149German Arena Text · german: Human pairwise preferences in the stated text vote pool Arena 1,468 #23 of 154instruction following Arena Text · instruction following: Human pairwise preferences in the stated text vote pool Arena 1,457 #31 of 188Japanese Arena Text · japanese: Human pairwise preferences in the stated text vote pool Arena 1,448 #20 of 138Korean Arena Text · korean: Human pairwise preferences in the stated text vote pool Arena 1,458 #13 of 144non-English Arena Text · non english: Human pairwise preferences in the stated text vote pool Arena 1,471 #15 of 188Polish Arena Text · polish: Human pairwise preferences in the stated text vote pool Arena 1,495 #6 of 133Russian Arena Text · russian: Human pairwise preferences in the stated text vote pool Arena 1,482 #15 of 186Spanish Arena Text · spanish: Human pairwise preferences in the stated text vote pool Arena 1,465 #24 of 155overall · full cohort Arena Text · overall · snapshot 2026-10-15 · native board settings: Human pairwise preferences in the stated text vote pool Arena 1,478 #15 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 1,435 #38 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 70.6% #39 of 95
EQ-Bench Creative Writingv3 Creative Writing v3 · Elo: Judge-rated creative writing under the v3 prompt set and pairwise ranking EQ-Bench 1,927 #12 of 91
Lech Mazur Short-Story Lech Mazur Short-Story · comparison score: Judges compare stories written to ten required elements; the comparison score can be negative Lech 0.8 #16 of 44
Health & law 3
Vals LegalBench Vals LegalBench: Legal reasoning and legal-language understanding on LegalBench Vals 85% #23 of 89
Vals MedScribe Vals MedScribe: Accuracy of clinical notes produced from consultation transcripts Vals 88.9% #9 of 80
Vals Legal Research Bench Vals Legal Research Bench: Grounded legal research; a task passes only if all criteria pass Vals 38% #32 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 77.7% #73 of 275
GDP.pdfAA all-pass GDP.pdf · AA all-pass: Models reason over professional PDF documents; an attempt passes only when every atomic criterion passes 14.4% #41 of 133
Arena Document Arena Document · overall: Human pairwise preferences in the stated document vote pool Arena 1,465 #15 of 33
Legacy · saturated or no longer reported 2
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
SWE-bench Pro (public) SWE-bench Pro (public): Agents resolve long-horizon professional repository tasks; public and private splits are separate 61.5% #1 of 19
No benchmark matches
CheaperGPT-6 Solmax×2.2-24%$1.04
OpenKimi K3max+9%+45%$2.00Sampling parameters are fixed+3
CheaperGemini 3.8 Flashhigh+68%-10%*$1.24Sampling parameters are ignored+1
Qwen3.8 FlashxhighOpen-29%-89%$0.15 / $0.47Context: 1M → 262K+6
GLM 5.3 FlashmaxOpen-33%-79%$0.28Defaults to max · Self-hosted+4