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 10 #208 of 337
Knowledge 6
AA OmniscienceAccuracy Measures correct answers on the AA Omniscience factual knowledge test AA 15.1% #234 of 321+1 versions & slices Hallucination Rate Measures how often attempted answers are incorrect on the AA Omniscience factual knowledge test AA 86.7% #220 of 321
AA Omniscience Index Scores factual knowledge while penalizing incorrect answers to encourage admitting uncertainty AA -58.6 #250 of 321
IFBench (instructions) Checks whether the model follows precise formatting and content instructions AA 57.8% #91 of 248
EQ-Bench3 Judge-rated responses to social and emotionally nuanced scenarios, scored through the source pairwise ranking EQ-Bench 956 #40 of 50
SimpleQA Verifiedv1 Accuracy on short factual questions with verified answers, answered without web access 5.9% #82 of 83
Math 5
MathArenaAIME 2025 Answer accuracy on the named mathematics competition under the recorded effort MathArena 89.2% #12 of 30+4 versions & slices BRUMO 2025 Answer accuracy on the named mathematics competition under the recorded effort MathArena 86.7% #19 of 24CMIMC 2025 Answer accuracy on the named mathematics competition under the recorded effort MathArena 73.8% #14 of 22HMMT February 2025 Answer accuracy on the named mathematics competition under the recorded effort MathArena 76.7% #16 of 30SMT 2025 Answer accuracy on the named mathematics competition under the recorded effort MathArena 83% #17 of 24
Science 3
Humanity's Last Exam Very hard expert questions across many fields, written to stump today’s models AA 5.3% #266 of 337
CritPt (physics) Research-level physics problems that need many careful reasoning steps AA 0% #179 of 324
SciCode Writing code for real scientific computing problems (physics, biology, math) AA 38.9% #101 of 140
Coding 3
Terminal-Bench4.0 Agents use a terminal to complete executable coding, system administration and research tasks in the 4.0 suite 0% #75 of 132+1 versions & slices 2.1 Older, easier version of Terminal-Bench; Still widely reported AA 13.9% #119 of 156
Terminal-Bench Hard Agents solve difficult software, system-administration and data-processing tasks in a terminal AA 4.5% #162 of 242
Agents 5
AutomationBench-AA Agents orchestrate SaaS workflows through APIs; objectives receive credit only when the task has no guardrail violation AA 0.2% #131 of 133
AA-Briefcase1.1 · Elo Agents produce professional deliverables; blind pairwise comparisons produce the source Elo rating AA 0 #129 of 140
APEX-Agents Agents complete long-horizon professional-services tasks across realistic applications AA 0.7% #29 of 29
GDPval-AA (work tasks)2.1 Real work tasks from many professions (reports, spreadsheets, analyses), graded against expert work AA 0 #122 of 175
τ³-BankingAA Agents retrieve banking knowledge and execute tools to resolve customer-support workflows AA 7% #103 of 134
Writing 16
Arena Textcreative writing Human pairwise preferences in the stated text vote pool Arena 1,240 #165 of 195+13 versions & slices overall Human pairwise preferences in the stated text vote pool Arena 1,317 #149 of 196Chinese Human pairwise preferences in the stated text vote pool Arena 1,315 #135 of 181English Human pairwise preferences in the stated text vote pool Arena 1,303 #147 of 188German Human pairwise preferences in the stated text vote pool Arena 1,255 #123 of 154instruction following Human pairwise preferences in the stated text vote pool Arena 1,236 #155 of 188Japanese Human pairwise preferences in the stated text vote pool Arena 1,244 #101 of 138Korean Human pairwise preferences in the stated text vote pool Arena 1,236 #113 of 144non-English Human pairwise preferences in the stated text vote pool Arena 1,269 #144 of 188Polish Human pairwise preferences in the stated text vote pool Arena 1,252 #122 of 133Russian Human pairwise preferences in the stated text vote pool Arena 1,278 #139 of 186Spanish Human pairwise preferences in the stated text vote pool Arena 1,267 #133 of 155overall · full cohort Human pairwise preferences in the stated text vote pool Arena 1,287 #146 of 188creative writing · full cohort Human pairwise preferences in the stated text vote pool Arena 1,201 #161 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 21.2% #94 of 95
EQ-Bench Creative Writingv3 Creative Writing v3 · Elo: Judge-rated creative writing under the v3 prompt set and pairwise ranking EQ-Bench 666 #89 of 91
Health & law 2
MLCR-AA Models synthesize long medical records; the headline credits concise answers judged both accurate and complete AA 0.6% #60 of 64
Vals LegalBench Legal reasoning and legal-language understanding on LegalBench Vals 70.9% #81 of 89
Documents 2
GDP.pdfAA all-pass Models reason over professional PDF documents; an attempt passes only when every atomic criterion passes 2% #106 of 133
AA-LCR (long context)1.1 Questions that need reading and connecting facts across very long documents (around 100K tokens) AA 31% #220 of 275
Legacy · saturated or no longer reported 13
AIME2025 · AA archived AA evaluates the 30 integer-answer problems from the 2025 AIME mathematics competition AA 62.3% #77 of 140
LiveCodeBenchAA archived AA evaluates solutions to newly collected competitive-programming problems AA 65.2% #62 of 153
GPQA Diamond Graduate-level biology, chemistry and physics questions that can’t be solved by a quick web search 61.1% #225 of 315+1 versions & slices Epoch Models answer graduate-level science multiple-choice questions Epoch 53.2% #99 of 133
MMLU-Proofficial board Graduate-level multiple-choice knowledge and reasoning across 14 subjects, including science, mathematics, law and health 73.1% #3 of 32
OTIS Mock AIME 20242025 Models solve the stated OTIS mock AIME mathematics problems Epoch 40.3% #96 of 130
SWE-rebenchall · 2025-08-01 to 2025-09-01 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 8.1% #3 of 4+5 versions & slices Go · 2025-08-01 to 2025-09-01 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 4Java · 2025-08-01 to 2025-09-01 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 4Python · 2025-08-01 to 2025-09-01 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 8.1% #3 of 4Rust · 2025-08-01 to 2025-09-01 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 4TypeScript · 2025-08-01 to 2025-09-01 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 50.3% #110 of 245
No benchmark matches
OpenQwen3.8 27Bunstatedwins 70%+2,556%$0.42 / $2.55Memory (4-bit): 13 GB → 17 GB+5
Muse Spark 1.2xhighwins 80%+5,456%$0.97Only auto tool_choice is supported+1
Gemma 3 27BunstatedOpenwins 60%+178%$0.08 / $0.16Memory (4-bit): 13 GB → 16 GB