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 4.9 #359 of 373
Knowledge 6
AA OmniscienceAccuracy Measures correct answers on the AA Omniscience factual knowledge test AA 13% #271 of 339+1 versions & slices Hallucination Rate Measures how often attempted answers are incorrect on the AA Omniscience factual knowledge test AA 92.1% #291 of 339
AA Omniscience Index Scores factual knowledge while penalizing incorrect answers to encourage admitting uncertainty AA -67.2 #303 of 339
IFBench (instructions) Checks whether the model follows precise formatting and content instructions AA 31.8% #215 of 271
EQ-Bench3 Judge-rated responses to social and emotionally nuanced scenarios, scored through the source pairwise ranking EQ-Bench 1,105 #36 of 53
SimpleQA Verifiedv1 Accuracy on short factual questions with verified answers, answered without web access 10.7% #78 of 84
Math 1
MATH Level5 Models solve competition mathematics problems from the hardest MATH difficulty level Epoch 74% #11 of 35
Science 3
Humanity's Last Exam Very hard expert questions across many fields, written to stump today’s models AA 4.4% #321 of 370
CritPt (physics) Research-level physics problems that need many careful reasoning steps AA 0% #184 of 342
SciCode Writing code for real scientific computing problems (physics, biology, math) AA 23.3% #137 of 146
Coding 3
Terminal-Bench4.0 Agents use a terminal to complete executable coding, system administration and research tasks in the 4.0 suite 0% #78 of 138+1 versions & slices 2.1 Older, easier version of Terminal-Bench; Still widely reported AA 4.5% #140 of 161
Terminal-Bench Hard Agents solve difficult software, system-administration and data-processing tasks in a terminal AA 3.8% #176 of 258
Agents 3
AutomationBench-AA Agents orchestrate SaaS workflows through APIs; objectives receive credit only when the task has no guardrail violation AA 0.2% #135 of 139
AA-Briefcase1.1 · Elo Agents produce professional deliverables; blind pairwise comparisons produce the source Elo rating AA 0 #133 of 146
τ³-BankingAA Agents retrieve banking knowledge and execute tools to resolve customer-support workflows AA 0.8% #135 of 137
Multimodal 3
MMMU-Pro (vision) College-level questions that require understanding images: charts, diagrams, photos AA 48% #118 of 136
MMMUvalidation Image-grounded university-level questions across 30 disciplines; validation and test sets remain separate MMMU 64.9% #19 of 41
Arena Vision Human pairwise preferences in the stated vision vote pool Arena 1,165 #70 of 88
Writing 19
Arena Textcreative writing · October 22, 2026 Human pairwise preferences in the stated text vote pool Arena 1,345 #100 of 197+16 versions & slices creative writing Human pairwise preferences in the stated text vote pool Arena 1,348 #104 of 195overall Human pairwise preferences in the stated text vote pool Arena 1,365 #115 of 196Chinese Human pairwise preferences in the stated text vote pool Arena 1,346 #140 of 198English Human pairwise preferences in the stated text vote pool Arena 1,376 #123 of 205French Human pairwise preferences in the stated text vote pool Arena 1,368 #113 of 164German Human pairwise preferences in the stated text vote pool Arena 1,362 #97 of 167instruction following Human pairwise preferences in the stated text vote pool Arena 1,321 #129 of 205Japanese Human pairwise preferences in the stated text vote pool Arena 1,287 #100 of 152Korean Human pairwise preferences in the stated text vote pool Arena 1,307 #100 of 156non-English Human pairwise preferences in the stated text vote pool Arena 1,334 #124 of 205Polish Human pairwise preferences in the stated text vote pool Arena 1,345 #100 of 142Russian Human pairwise preferences in the stated text vote pool Arena 1,349 #118 of 203Spanish Human pairwise preferences in the stated text vote pool Arena 1,349 #124 of 168overall · October 15, 2026 Human pairwise preferences in the stated text vote pool Arena 1,358 #114 of 188creative writing · October 15, 2026 Human pairwise preferences in the stated text vote pool Arena 1,345 #96 of 187overall · October 22, 2026 Human pairwise preferences in the stated text vote pool Arena 1,358 #117 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 41.9% #75 of 96
EQ-Bench Creative Writingv3 Creative Writing v3 · Elo: Judge-rated creative writing under the v3 prompt set and pairwise ranking EQ-Bench 1,266 #71 of 94
Documents 2
AA-LCR (long context)1.1 Questions that need reading and connecting facts across very long documents (around 100K tokens) AA 7.3% #279 of 338
GDP.pdfAA all-pass Models reason over professional PDF documents; an attempt passes only when every atomic criterion passes 0.2% #135 of 139
Legacy · saturated or no longer reported 15
AIME2025 · AA archived AA evaluates the 30 integer-answer problems from the 2025 AIME mathematics competition AA 20.7% #124 of 161
LiveCodeBenchAA archived AA evaluates solutions to newly collected competitive-programming problems AA 13.7% #161 of 184
GPQA Diamond Graduate-level biology, chemistry and physics questions that can’t be solved by a quick web search 42.8% #285 of 346+1 versions & slices Epoch Models answer graduate-level science multiple-choice questions Epoch 47.7% #107 of 134
Aider Polyglottwo attempts Models edit code for 225 Exercism exercises in six languages, with a second attempt after test feedback Aider 4.9% #14 of 15
BFCLv4 Tests function calling, multi-turn interactions, search and memory in the v4 suite Berkeley 29.5% #25 of 39
SWE-bench Pro (public) Agents resolve long-horizon professional repository tasks; public and private splits are separate 11.4% #17 of 19
OTIS Mock AIME 20242025 Models solve the stated OTIS mock AIME mathematics problems Epoch 22.5% #103 of 131
SWE-rebenchall · January 1, 2025 to September 1, 2025 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 5.4% #4 of 4+5 versions & slices Go · January 1, 2025 to September 1, 2025 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 4Java · January 1, 2025 to September 1, 2025 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 4Python · January 1, 2025 to September 1, 2025 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 5.4% #4 of 4Rust · January 1, 2025 to September 1, 2025 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 4TypeScript · January 1, 2025 to September 1, 2025 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 4
τ²-benchAA Tests service agents that use tools while coordinating with a simulated user AA 10.5% #238 of 262
No benchmark matches
CheaperGPT OSS 120BunstatedOpensame-40%$0.03 / $0.15Memory (4-bit): 16 GB → 70 GB+1
OpenQwen3.8 27Bunstatedwins 63%+856%$0.42 / $2.55Memory (4-bit): 16 GB → 17 GB+5
BetterClaude Haiku 5.5highwins 60%+100%$0.079Remove sampling parameters+5