Q1 How high is its intelligence index across knowledge work, coding and science? AA Intelligence Index v5 · Bottom 28%#40 Q1 How high is its intelligence index across knowledge work, coding and science? AA Intelligence Index v5 · Bottom 28%#40 Q2 Which AI gives the answers people prefer most? Arena Text · Bottom 29%#78 Q3 On the old v1, could it solve problems it had not seen in training? LiveBench Overall v1 · Bottom 10%#48
Q1 Bottom 28%#40 of 54 18.9 Q2 1434 Q3 67.3 About AA Intelligence Index v5 More about AA Intelligence Index v5
Q1 Can it handle customer service by the rules? τ³-Banking · Bottom 38%#50 Q1 Can it handle customer service by the rules? τ³-Banking · Bottom 38%#50 Q2 Can it run a small business alone for a year and make money? Vending-Bench 2 · Bottom 20%#47 Q3 Can it handle customer service by the rules? TAU2 · Top 49%#38 Q1 Bottom 38%#50 of 79 15.3% Q2 $2,158 Q3 88.9% Q4 74.2% Q5 22.3% About τ³-Banking More about τ³-Banking
Q1 Can it pull scattered facts from several long documents into one answer? Long Context Reasoning · Top 14%#16 Q1 Can it pull scattered facts from several long documents into one answer? Long Context Reasoning · Top 14%#16 Q2 Which AI gives the business answers people prefer most? Arena Business, Management & Finance · Bottom 38%#68 Q3 Can it do professional work in banking, consulting and law? APEX-Agents · Bottom 17%#50 Q1 Top 14%#16 of 121 83.0% Q2 1443 Q3 37.7% Q4 48.3% Q5 68.1% Q6 20.6% Q7 29.8% Q8 46.3% Q9 1437 Q10 76.2 Q11 54.3% Q12 4.2% About Long Context Reasoning More about Long Context Reasoning
Q1 Which AI gives answers people prefer most, without false claims? Arena Text Factuality · Bottom 25%#79 Q1 Which AI gives answers people prefer most, without false claims? Arena Text Factuality · Bottom 25%#79 Q2 Can it track down hard-to-find information on the web? BrowseComp · Top 38%#15
Q1 Bottom 25%#79 of 103 1433 Q2 83.5% About Arena Text Factuality More about Arena Text Factuality
Q1 Can AI carry out long development tasks autonomously by operating a computer itself? Terminal-Bench 4.0 · Bottom 47%#46 Q1 Can AI carry out long development tasks autonomously by operating a computer itself? Terminal-Bench 4.0 · Bottom 47%#46 Q2 Which AI do people prefer most for coding answers? Arena Coding · Bottom 35%#72 Q3 Can AI test and fix code itself until it completes the given task? LiveBench Agentic Coding v1 · Bottom 12%#47 Q1 Bottom 47%#46 of 84 2.0% Q2 1492 Q3 40.7 Q4 640 Q5 59.0% Q6 65.2% Q7 42.4% Q8 58.6 Q9 1484 Q10 47.1% Q11 80.5% Q12 14.7% Q13 1482 Q14 68.2 Q15 47.6% Q16 0.9% Q17 52.6% About Terminal-Bench 4.0 More about Terminal-Bench 4.0
Q1 Which AI gives the writing and literature answers people prefer most? Arena Writing, Literature & Language · Bottom 27%#80 Q1 Which AI gives the writing and literature answers people prefer most? Arena Writing, Literature & Language · Bottom 27%#80 Q2 Which AI do people prefer most in long conversations? Arena Multi-turn · Bottom 36%#71 Q3 Does it follow tricky rules like format and length without missing any? IFBench · Top 3%#2 Q1 Bottom 27%#80 of 108 1410 Q2 1447 Q3 82.9% Q4 1392 Q5 1426 Q6 76.8 Q7 1140 Q8 87.3% Q9 1398 Q10 57.5 About Arena Writing, Literature & Language More about Arena Writing, Literature & Language
Q1 Can it solve problems that stump experts across fields? Humanity's Last Exam · Top 38%#46 Q1 Can it solve problems that stump experts across fields? Humanity's Last Exam · Top 38%#46 Q2 Can it solve research-level physics problems? CritPt · Bottom 33%#60 Q3 Which AI gives the math answers people prefer most? Arena Math · Bottom 22%#83 Q1 Top 38%#46 of 122 39.0% Q2 3.7% Q3 1423 Q4 92.9% Q5 71.1% Q6 77.0 Q7 18.0% Q8 55.0% Q9 1458 Q10 1442 Q11 74.5 Q12 45.8% Q13 17.0% Q14 8.0% Q15 78.9% Q16 33.7 About Humanity's Last Exam More about Humanity's Last Exam
Q1 Which AI gives the photo and chart answers people prefer most? Arena Vision · Bottom 35%#44 Q1 Which AI gives the photo and chart answers people prefer most? Arena Vision · Bottom 35%#44 Q2 Can it solve college-level problems with images and diagrams? MMMU-Pro · Top 46%#17
Q1 Bottom 35%#44 of 66 1248 Q2 78.1% About Arena Vision More about Arena Vision
Q1 Which AI do people prefer most when asking in Korean? Arena Korean · Bottom 17%#77 Q1 Which AI do people prefer most when asking in Korean? Arena Korean · Bottom 17%#77 Q2 Which AI do people prefer most when asking in Japanese? Arena Japanese · Bottom 20%#72 Q3 Which AI do people prefer most on tough questions? Arena Hard Prompts · Bottom 28%#79 Q1 Bottom 17%#77 of 91 1361 Q2 1366 Q3 1456 Q4 1467 Q5 1450 Q6 1432 Q7 97.6% Q8 82.7% Q9 43.8% Q10 80.6% Q11 85.3% Q12 79.0% Q13 75.2% About Arena Korean More about Arena Korean
Q1 Does it call out a nonsense question instead of playing along? BullshitBench v2 · Top 25%#35 Q1 Does it call out a nonsense question instead of playing along? BullshitBench v2 · Top 25%#35 Q2 Can it sort word-puzzle tiles into groups using meaning and wordplay? NYT Connections Extended · Bottom 40%#71 Q3 Who survives longer in head-to-head snake games? SnakeBench · Top 38%#112 Q4 What would it score on Korea's 2026 CSAT (Suneung)? CSAT 2026 · Bottom 25%#47
Q1 Top 25%#35 of 143 64.0% Reasoning effort: Extra High Q2 65.1% Q3 25 Q4 330 By effort None 63.0% Extra High 64.0%
About BullshitBench v2 More about BullshitBench v2