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 8.5 #240 of 339
Knowledge 5
AA OmniscienceAccuracy Measures correct answers on the AA Omniscience factual knowledge test AA 25.5% #110 of 323+1 versions & slices Hallucination Rate Measures how often attempted answers are incorrect on the AA Omniscience factual knowledge test AA 90% #253 of 323
AA Omniscience Index Scores factual knowledge while penalizing incorrect answers to encourage admitting uncertainty AA -41.7 #182 of 323
IFBench (instructions) Checks whether the model follows precise formatting and content instructions AA 34.8% #186 of 248
SimpleQA Verifiedv1 Accuracy on short factual questions with verified answers, answered without web access 27.4% #63 of 84
Math 2
FrontierMath Models solve expert-written advanced mathematics problems on the named tier and edition Epoch 1.7% #36 of 40
MATH Level5 Models solve competition mathematics problems from the hardest MATH difficulty level Epoch 64.9% #16 of 35
Science 3
Humanity's Last Exam Very hard expert questions across many fields, written to stump today’s models AA 2.9% #334 of 339
CritPt (physics) Research-level physics problems that need many careful reasoning steps AA 0% #182 of 326
SciCode Writing code for real scientific computing problems (physics, biology, math) AA 35.8% #118 of 142
Coding 9
Terminal-Bench4.0 Agents use a terminal to complete executable coding, system administration and research tasks in the 4.0 suite 0% #77 of 134+1 versions & slices 2.1 Older, easier version of Terminal-Bench; Still widely reported AA 16.9% #116 of 156
SWE-rebenchall · January 1, 2025 to August 1, 2025 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 23.3% #4 of 6+5 versions & slices Go · January 1, 2025 to August 1, 2025 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 6Java · January 1, 2025 to August 1, 2025 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 6Python · January 1, 2025 to August 1, 2025 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 23.3% #4 of 6Rust · January 1, 2025 to August 1, 2025 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 6TypeScript · January 1, 2025 to August 1, 2025 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 6
Terminal-Bench Hard Agents solve difficult software, system-administration and data-processing tasks in a terminal AA 6.8% #144 of 242
Agents 4
AutomationBench-AA Agents orchestrate SaaS workflows through APIs; objectives receive credit only when the task has no guardrail violation AA 0.6% #119 of 135
AA-Briefcase1.1 · Elo Agents produce professional deliverables; blind pairwise comparisons produce the source Elo rating AA 77.5 #125 of 142
GDPval-AA (work tasks)2.1 Real work tasks from many professions (reports, spreadsheets, analyses), graded against expert work AA -29.3 #172 of 177
τ³-BankingAA Agents retrieve banking knowledge and execute tools to resolve customer-support workflows AA 4.7% #123 of 134
Writing 16
Arena Textcreative writing · October 22, 2026 Human pairwise preferences in the stated text vote pool Arena 1,329 #113 of 197+15 versions & slices creative writing Human pairwise preferences in the stated text vote pool Arena 1,349 #102 of 195overall Human pairwise preferences in the stated text vote pool Arena 1,358 #118 of 196Chinese Human pairwise preferences in the stated text vote pool Arena 1,338 #141 of 198English Human pairwise preferences in the stated text vote pool Arena 1,346 #145 of 205French Human pairwise preferences in the stated text vote pool Arena 1,343 #123 of 164German Human pairwise preferences in the stated text vote pool Arena 1,324 #117 of 167instruction following Human pairwise preferences in the stated text vote pool Arena 1,315 #134 of 205Japanese Human pairwise preferences in the stated text vote pool Arena 1,266 #106 of 152Korean Human pairwise preferences in the stated text vote pool Arena 1,248 #119 of 156non-English Human pairwise preferences in the stated text vote pool Arena 1,315 #136 of 205Russian Human pairwise preferences in the stated text vote pool Arena 1,324 #132 of 203Spanish Human pairwise preferences in the stated text vote pool Arena 1,352 #122 of 168overall · October 15, 2026 Human pairwise preferences in the stated text vote pool Arena 1,332 #130 of 188creative writing · October 15, 2026 Human pairwise preferences in the stated text vote pool Arena 1,329 #107 of 187overall · October 22, 2026 Human pairwise preferences in the stated text vote pool Arena 1,332 #133 of 198
Health & law 1
Vals LegalBench Legal reasoning and legal-language understanding on LegalBench Vals 80.8% #61 of 89
Documents 2
GDP.pdfAA all-pass Models reason over professional PDF documents; an attempt passes only when every atomic criterion passes 0.8% #124 of 135
AA-LCR (long context)1.1 Questions that need reading and connecting facts across very long documents (around 100K tokens) AA 29.3% #223 of 317
Legacy · saturated or no longer reported 11
AIME2025 · AA archived AA evaluates the 30 integer-answer problems from the 2025 AIME mathematics competition AA 26% #109 of 140
LiveCodeBenchAA archived AA evaluates solutions to newly collected competitive-programming problems AA 35.9% #109 of 153+1 versions & slices official artifact (through April 2025) Mean pass@1 over each model’s available problems in the current official public generation artifact; its latest problem is dated April 7, 2025 LiveCodeBench 49.6% #8 of 10
GPQA Diamond Graduate-level biology, chemistry and physics questions that can’t be solved by a quick web search 55.7% #246 of 315+2 versions & slices Epoch Models answer graduate-level science multiple-choice questions Epoch 71.2% #80 of 134January 2025 run Models answer graduate-level science multiple-choice questions Epoch 56.5% #1 of 1
OTIS Mock AIME 20242025 Models solve the stated OTIS mock AIME mathematics problems Epoch 48.9% #94 of 131+1 versions & slices 2025 · February 2025 run Models solve the stated OTIS mock AIME mathematics problems Epoch 15.8% #1 of 1
MathArenaAIME 2025 Answer accuracy on the named mathematics competition under the recorded effort MathArena 25% #29 of 30+1 versions & slices HMMT February 2025 Answer accuracy on the named mathematics competition under the recorded effort MathArena 13.3% #29 of 30
τ²-benchAA Tests service agents that use tools while coordinating with a simulated user AA 22.8% #180 of 245
No benchmark matches
CheaperGLM 5.3 FlashunstatedOpenwins 68%-25%$0.15 / $0.50Defaults to max · Self-hosted+4
BetterGPT-5 Minihighwins 56%+118%$0.053
CheaperMiMo v2.6 FlashunstatedOpenwins 68%-44%$0.14 / $0.28Pass back reasoning_content+3
CheaperHy3unstatedOpenwins 67%-27%$0.13 / $0.53
Qwen3 235B A22B Instruct 2507unstatedOpenwins 65%-35%$0.09 / $0.55
NVIDIA Nemotron 3 Ultra 550B A55B NVFP4Openwins 70%——