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 9.7 #214 of 339
Knowledge 5
AA OmniscienceAccuracy Measures correct answers on the AA Omniscience factual knowledge test AA 24.3% #123 of 323+1 versions & slices Hallucination Rate Measures how often attempted answers are incorrect on the AA Omniscience factual knowledge test AA 85.9% #213 of 323
AA Omniscience Index Scores factual knowledge while penalizing incorrect answers to encourage admitting uncertainty AA -40.7 #178 of 323
IFBench (instructions) Checks whether the model follows precise formatting and content instructions AA 41% #155 of 248
EQ-Bench3 Judge-rated responses to social and emotionally nuanced scenarios, scored through the source pairwise ranking EQ-Bench 1,124 #34 of 53
Math 1
MATH Level5 Models solve competition mathematics problems from the hardest MATH difficulty level Epoch 75.5% #10 of 35
Science 3
Humanity's Last Exam Very hard expert questions across many fields, written to stump today’s models AA 4.7% #291 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 39% #101 of 142
Coding 15
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 13.9% #119 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 24.6% #3 of 6+11 versions & slices all · January 1, 2025 to November 1, 2025 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 24.8% #1 of 1Go · 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 6Go · January 1, 2025 to November 1, 2025 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 1Java · 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 November 1, 2025 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 1Python · January 1, 2025 to August 1, 2025 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 24.6% #3 of 6Python · January 1, 2025 to November 1, 2025 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 24.8% #1 of 1Rust · 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 6Rust · January 1, 2025 to November 1, 2025 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 1TypeScript · 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 November 1, 2025 Repository issue resolution over the exact fresh pull-request window, under the source agent version Nebius 0% #1 of 1
Terminal-Bench Hard Agents solve difficult software, system-administration and data-processing tasks in a terminal AA 15.2% #109 of 242
Agents 4
AutomationBench-AA Agents orchestrate SaaS workflows through APIs; objectives receive credit only when the task has no guardrail violation AA 0.3% #129 of 135
AA-Briefcase1.1 · Elo Agents produce professional deliverables; blind pairwise comparisons produce the source Elo rating AA 68 #128 of 142
GDPval-AA (work tasks)2.1 Real work tasks from many professions (reports, spreadsheets, analyses), graded against expert work AA 68.8 #121 of 177
τ³-BankingAA Agents retrieve banking knowledge and execute tools to resolve customer-support workflows AA 4.7% #123 of 134
Writing 15
Arena Textcreative writing · October 15, 2026 Human pairwise preferences in the stated text vote pool Arena 1,364 #85 of 187+12 versions & slices Chinese Human pairwise preferences in the stated text vote pool Arena 1,390 #119 of 198English Human pairwise preferences in the stated text vote pool Arena 1,389 #112 of 205French Human pairwise preferences in the stated text vote pool Arena 1,385 #106 of 164German Human pairwise preferences in the stated text vote pool Arena 1,374 #88 of 167instruction following Human pairwise preferences in the stated text vote pool Arena 1,344 #118 of 205Japanese Human pairwise preferences in the stated text vote pool Arena 1,333 #85 of 152Korean Human pairwise preferences in the stated text vote pool Arena 1,319 #96 of 156non-English Human pairwise preferences in the stated text vote pool Arena 1,358 #115 of 205Polish Human pairwise preferences in the stated text vote pool Arena 1,368 #91 of 142Russian Human pairwise preferences in the stated text vote pool Arena 1,373 #105 of 203Spanish Human pairwise preferences in the stated text vote pool Arena 1,358 #117 of 168overall · October 15, 2026 Human pairwise preferences in the stated text vote pool Arena 1,375 #103 of 188
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 60.1% #52 of 96
EQ-Bench Creative Writingv3 Creative Writing v3 · Elo: Judge-rated creative writing under the v3 prompt set and pairwise ranking EQ-Bench 1,472 #55 of 94
Documents 2
GDP.pdfAA all-pass Models reason over professional PDF documents; an attempt passes only when every atomic criterion passes 1.6% #111 of 135
AA-LCR (long context)1.1 Questions that need reading and connecting facts across very long documents (around 100K tokens) AA 40.7% #200 of 317
Legacy · saturated or no longer reported 8
AIME2025 · AA archived AA evaluates the 30 integer-answer problems from the 2025 AIME mathematics competition AA 41% #95 of 140
LiveCodeBenchAA archived AA evaluates solutions to newly collected competitive-programming problems AA 40.5% #104 of 153
GPQA Diamond Graduate-level biology, chemistry and physics questions that can’t be solved by a quick web search 65.5% #213 of 315+1 versions & slices Epoch Models answer graduate-level science multiple-choice questions Epoch 67.6% #84 of 134
OTIS Mock AIME 20242025 Models solve the stated OTIS mock AIME mathematics problems Epoch 37.8% #99 of 131
MathArenaAIME 2025 Answer accuracy on the named mathematics competition under the recorded effort MathArena 50% #27 of 30+1 versions & slices HMMT February 2025 Answer accuracy on the named mathematics competition under the recorded effort MathArena 29.2% #27 of 30
τ²-benchAA Tests service agents that use tools while coordinating with a simulated user AA 47.1% #115 of 245
No benchmark matches
CheaperNemotron 3 Nano 30B A3BthinkingOpen×2.2-18%$0.017
BetterGPT OSS 20BhighOpen×2.2-49%$0.011
CheaperSeed OSS 36B InstructthinkingOpen×2.1-26%$0.21 / $0.57
Qwen3 30B A3B 2507 (Reasoning)thinkingOpen+37%+222%$0.066
GLM 4.7thinkingOpen×2.3+147%$0.60 / $2.20
Kimi K2 ThinkingthinkingOpen×2.3+110%$0.47 / $2Memory (4-bit): 411 GB → 616 GB