Q1 Which AI gives the answers people prefer most? Arena Text · Top 11%#11 Q1 Which AI gives the answers people prefer most? Arena Text · Top 11%#11 Q2 On the old v1, could it solve problems it had not seen in training? LiveBench Overall v1 · Bottom 45%#30 Q3 How high was its intelligence index on the old v4 edition? AA Intelligence Index v4 · Bottom 50%#39
Q1 Top 11%#11 of 108 1496 Q2 74.5 Q3 31.9 About Arena Text More about Arena Text
Q1 Can it run a small business alone for a year and make money? Vending-Bench 2 · Top 34%#19 Q1 Can it run a small business alone for a year and make money? Vending-Bench 2 · Top 34%#19 Q2 Can it handle customer service by the rules? TAU2 · Top 40%#31 Q3 Can it pick the right tools and use them properly? MCP-Atlas · Top 44%#16 Q1 Top 34%#19 of 57 $8,018 Q2 92.1% Q3 75.8% Q4 4.2% Q5 11h 59m Q6 57.7% About Vending-Bench 2 More about Vending-Bench 2
Q1 Can it pull scattered facts from several long documents into one answer? Long Context Reasoning · Top 48%#57 Q1 Can it pull scattered facts from several long documents into one answer? Long Context Reasoning · Top 48%#57 Q2 Which AI gives the business answers people prefer most? Arena Business, Management & Finance · Top 6%#6 Q3 Can it do professional work in banking, consulting and law? APEX-Agents · Bottom 31%#42 Q1 Top 48%#57 of 121 78.0% Q2 1500 Q3 46.3% Q4 67.0% Q5 49.1% Q6 1508 Q7 69.9 Q8 60.0% Q9 60.7% Q10 20.7% Q11 17.0% By effort High (no thinking) 67.0% Max 78.0%
About Long Context Reasoning More about Long Context Reasoning
Q1 Which AI gives answers people prefer most, without false claims? Arena Text Factuality · Top 7%#7 Q1 Which AI gives answers people prefer most, without false claims? Arena Text Factuality · Top 7%#7 Q2 Does it know facts accurately without searching? SimpleQA Verified · Bottom 50%#36 Q3 Can it carry out multi-step web research? Deep Research Bench · Top 6%#1 Q1 Top 7%#7 of 103 1493 Reasoning effort: High Q2 47.0% Q3 55.3% Q4 86.6% Q5 53.0% Q6 1252 Q7 12.2% Q8 87.8% Q9 59.9 Q10 92.5% About Arena Text Factuality More about Arena Text Factuality
Q1 Which AI do people prefer most for coding answers? Arena Coding · Top 8%#8 Q1 Which AI do people prefer most for coding answers? Arena Coding · Top 8%#8 Q2 Can AI test and fix code itself until it completes the given task? LiveBench Agentic Coding v1 · Bottom 41%#32 Q3 Can it find better answers to optimization problems with no single right answer? ALE-Bench · Bottom 46%#50 Q1 Top 8%#8 of 108 1547 Q2 49.0 Q3 997 Q4 53.4% Q5 48.5% Q6 48.1 Q7 1537 Q8 51.9% Q9 78.7% Q10 26.6% Q11 78.0% Q12 41.2% Q13 1534 Q14 78.2 Q15 57.6% Q16 79.8% Q17 33.3% Q18 44.1% Q19 1.47× About Arena Coding More about Arena Coding
Q1 Which AI gives the writing and literature answers people prefer most? Arena Writing, Literature & Language · Top 11%#11 Q1 Which AI gives the writing and literature answers people prefer most? Arena Writing, Literature & Language · Top 11%#11 Q2 Which AI do people prefer most in long conversations? Arena Multi-turn · Top 7%#7 Q3 Does it follow tricky rules like format and length without missing any? IFBench · Bottom 20%#66 Q1 Top 11%#11 of 108 1489 Q2 1509 Q3 53.1% Q4 1478 Q5 1498 Q6 83.3 Q7 1210 Q8 86.7% Q9 1476 Q10 63.3 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 35%#42 Q1 Can it solve problems that stump experts across fields? Humanity's Last Exam · Top 35%#42 Q2 Which AI gives the math answers people prefer most? Arena Math · Top 13%#13 Q3 Can it solve graduate-level science problems? GPQA Diamond · Top 36%#37 Q1 Top 35%#42 of 122 39.9% Q2 1504 Q3 89.6% Q4 94.4% Q5 89.3 Q6 66.0% Q7 26.8% Q8 50.0% Q9 1518 Q10 1511 Q11 88.7 Q12 40.7% Q13 22.9% Q14 94.0% Q15 69.2% Q16 67.6% Q17 7.6% Q18 21.0% Q19 25.0% Q20 91.2% Q21 55.8 By effort High (no thinking) 19.1% Max 39.9%
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 · Top 13%#8 Q1 Which AI gives the photo and chart answers people prefer most? Arena Vision · Top 13%#8 Q2 Can it spot the wrong step from assembly photos? Furniture Assembly · Bottom 22%#31 Q3 Can it read charts and answer accurately? CharXiv · Bottom 26%#24 Q4 Can it solve college-level problems with images and diagrams? MMMU-Pro · Bottom 30%#27
Q1 Top 13%#8 of 66 1312 Q2 28.3% Q3 77.4% Q4 73.9% About Arena Vision More about Arena Vision
Q1 Which AI do people prefer most when asking in Korean? Arena Korean · Top 17%#15 Q1 Which AI do people prefer most when asking in Korean? Arena Korean · Top 17%#15 Q2 Which AI do people prefer most when asking in Japanese? Arena Japanese · Top 21%#18 Q3 Can it use expert knowledge in many languages? MMMLU · Top 23%#4 Q1 Top 17%#15 of 91 1459 Q2 1479 Q3 91.1% Q4 1526 Q5 1534 Q6 1511 Q7 1508 Q8 93.0% Q9 66.6% Q10 12.7% Q11 96.7% Q12 54.6% Q13 36.8% Q14 76.7% Q15 36.8% Q16 96.2% Q17 70.2% Q18 70.2% Q19 97.6% Q20 85.0% Q21 50 Q22 211 Q23 264 Q24 -0.01 Q25 34.5% Q26 86.7% Q27 72.3% Q28 73.1% Q29 83.9% Q30 89.9% Q31 81.3% Q32 45.2% Q33 47.0% Q34 12.0% About Arena Korean More about Arena Korean
Q1 Does it call out a nonsense question instead of playing along? BullshitBench v2 · Top 5%#7 Q1 Does it call out a nonsense question instead of playing along? BullshitBench v2 · Top 5%#7 Q2 Can it sort word-puzzle tiles into groups using meaning and wordplay? NYT Connections Extended · Top 20%#22 Q3 Who survives longer in head-to-head snake games? SnakeBench · Top 6%#17 Q1 Top 5%#7 of 143 87.0% Reasoning effort: High Q2 92.1% Q3 33 Q4 570 Q5 892 Q6 324 By effort None 83.0% High 87.0%
About BullshitBench v2 More about BullshitBench v2