Q1 Which AI gives the answers people prefer most? Arena Text · Bottom 31%#76 Q1 Which AI gives the answers people prefer most? Arena Text · Bottom 31%#76 Q2 On the old v1, could it solve problems it had not seen in training? LiveBench Overall v1 · Bottom 6%#50 Q3 How high was its intelligence index on the old v4 edition? AA Intelligence Index v4 · Top 45%#34
Q1 Bottom 31%#76 of 108 1436 Q2 65.5 Q3 34.3 About Arena Text More about Arena Text
Q1 Can it handle customer service by the rules? τ³-Banking · Top 21%#16 Q1 Can it handle customer service by the rules? τ³-Banking · Top 21%#16 Q2 Can it handle customer service by the rules? TAU2 · Top 25%#19 Q3 Can it pick the right tools and use them properly? MCP-Atlas · Bottom 36%#25
Q1 Top 21%#16 of 79 39.4% Q2 95.0% Q3 69.0% About τ³-Banking More about τ³-Banking
Q1 Can it pull scattered facts from several long documents into one answer? Long Context Reasoning · Top 34%#40 Q1 Can it pull scattered facts from several long documents into one answer? Long Context Reasoning · Top 34%#40 Q2 Which AI gives the business answers people prefer most? Arena Business, Management & Finance · Bottom 30%#77 Q3 Can it find company filings and do financial analysis? Finance Agent v2 · Bottom 39%#41 Q1 Top 34%#40 of 121 79.7% Q2 1438 Q3 49.5% Q4 61.8% Q5 28.1% Q6 30.3% Q7 41.4% Q8 68.0 Q9 8.3% About Long Context Reasoning More about Long Context Reasoning
Q1 Which AI gives answers people prefer most, without false claims? Arena Text Factuality · Bottom 33%#71 Q1 Which AI gives answers people prefer most, without false claims? Arena Text Factuality · Bottom 33%#71 Q2 Does it know facts accurately without searching? SimpleQA Verified · Bottom 22%#55 Q3 Can it track down hard-to-find information on the web? BrowseComp · Bottom 35%#27 Q1 Bottom 33%#71 of 103 1439 Q2 33.6% Q3 73.2% Q4 45.1% Q5 4.3% 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 · Top 45%#37 Q1 Can AI carry out long development tasks autonomously by operating a computer itself? Terminal-Bench 4.0 · Top 45%#37 Q2 Which AI do people prefer most for coding answers? Arena Coding · Bottom 26%#81 Q3 Can AI test and fix code itself until it completes the given task? LiveBench Agentic Coding v1 · Bottom 4%#51 Q1 Top 45%#37 of 84 12.1% Q2 1484 Q3 37.6 Q4 1306 Q5 52.6% Q6 78.7% Q7 35.6% Q8 69.1 Q9 50.3% Q10 18.8% Q11 63.0% Q12 1471 Q13 69.2 Q14 0.0% Q15 74.7% 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 31%#76 Q1 Which AI gives the writing and literature answers people prefer most? Arena Writing, Literature & Language · Bottom 31%#76 Q2 Which AI do people prefer most in long conversations? Arena Multi-turn · Bottom 43%#63 Q3 Does it follow tricky rules like format and length without missing any? IFBench · Top 9%#7 Q1 Bottom 31%#76 of 108 1416 Q2 1453 Q3 79.2% Q4 1407 Q5 1428 Q6 70.1 Q7 80.4% Q8 1405 Q9 63.1 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 39%#47 Q1 Can it solve problems that stump experts across fields? Humanity's Last Exam · Top 39%#47 Q2 Can it solve research-level physics problems? CritPt · Top 43%#37 Q3 Which AI gives the math answers people prefer most? Arena Math · Bottom 24%#81 Q1 Top 39%#47 of 122 38.6% Q2 16.6% Q3 1424 Q4 90.8% Q5 94.4% Q6 79.7 Q7 57.5% Q8 24.4% Q9 56.0% Q10 88.4% Q11 1457 Q12 1434 Q13 70.6 Q14 89.0% Q15 61.4% Q16 61.1% Q17 32.0% Q18 34.0% Q19 90.9% Q20 41.7 About Humanity's Last Exam More about Humanity's Last Exam
Q1 Which AI do people prefer most when asking in Korean? Arena Korean · Bottom 33%#62 Q1 Which AI do people prefer most when asking in Korean? Arena Korean · Bottom 33%#62 Q2 Which AI do people prefer most when asking in Japanese? Arena Japanese · Bottom 32%#61 Q3 Which AI do people prefer most on tough questions? Arena Hard Prompts · Bottom 32%#75 Q1 Bottom 33%#62 of 91 1394 Q2 1397 Q3 1459 Q4 1457 Q5 1459 Q6 1448 Q7 95.8% Q8 93.9% Q9 94.9% Q10 85.1% Q11 27.1% Q12 89.4% Q13 58.9% Q14 81.7% Q15 86.0% Q16 80.0% Q17 70.7% About Arena Korean More about Arena Korean
Q1 Does it call out a nonsense question instead of playing along? BullshitBench v2 · Bottom 48%#76 Q1 Does it call out a nonsense question instead of playing along? BullshitBench v2 · Bottom 48%#76 Q2 Can it sort word-puzzle tiles into groups using meaning and wordplay? NYT Connections Extended · Top 26%#29 Q3 What would it score on Korea's 2026 CSAT (Suneung)? CSAT 2026 · Top 37%#22
Q1 Bottom 48%#76 of 143 39.0% Reasoning effort: Extra High Q2 89.6% Q3 414 By effort None 20.0% Extra High 39.0%
About BullshitBench v2 More about BullshitBench v2