Q1 How high is its intelligence index across knowledge work, coding and science? AA Intelligence Index v5 · Top 23%#12 Q1 How high is its intelligence index across knowledge work, coding and science? AA Intelligence Index v5 · Top 23%#12 Q2 Which AI gives the answers people prefer most? Arena Text · Top 2%#2 Q3 Can it solve problems it never trained on? LiveBench 2.0 · Top 31%#7 Q4 How high was its intelligence index on the old v4 edition? AA Intelligence Index v4 · Top 7%#5
Q1 Top 23%#12 of 54 50.4 Reasoning effort: High Q2 1520 Q3 77.7 Q4 52.6 By effort Medium 47.6 High 50.4
About AA Intelligence Index v5 More about AA Intelligence Index v5
Q1 Can it operate a computer to do professional work? Agents' Last Exam · Top 4%#1 Q1 Can it operate a computer to do professional work? Agents' Last Exam · Top 4%#1 Q2 Can it run a small business alone for a year and make money? Vending-Bench 2 · Top 11%#6 Q3 Would a client accept its work on freelance jobs? Remote Labor Index · Top 34%#5 Q1 Top 4%#1 of 26 39.5% Reasoning effort: Max Self-reported by developer Q2 $13,718 Q3 17.1% Q4 58.5 Q5 82.8 About Agents' Last Exam More about Agents' Last Exam
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 Can it read tables and reshape them as needed? LiveBench 2.0 Data Analysis · Top 5%#1 Q3 Which AI gives the business answers people prefer most? Arena Business, Management & Finance · Top 1%#1 Q1 Top 34%#40 of 121 79.7% Q2 85.2 Q3 1514 Q4 82.4% Q5 53.6% Q6 65.4% Q7 49.2% Q8 54.8% Q9 27.1% Q10 58.8% Q11 69.2 Q12 67.7% Q13 28.4% Q14 19.6% About Long Context Reasoning More about Long Context Reasoning
Q1 Which AI gives answers people prefer most, without false claims? Arena Text Factuality · Top 8%#8 Q1 Which AI gives answers people prefer most, without false claims? Arena Text Factuality · Top 8%#8 Q2 Does it know facts accurately without searching? SimpleQA Verified · Top 6%#4 Q3 Does it avoid making things up when summarizing? Vectara Hallucination Rate · Bottom 20%#33 Q1 Top 8%#8 of 103 1492 Q2 76.0% Q3 12.1% Q4 86.8 Q5 44.3% Q6 87.9% 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 6%#5 Q1 Can AI carry out long development tasks autonomously by operating a computer itself? Terminal-Bench 4.0 · Top 6%#5 Q2 Which AI do people prefer most for coding answers? Arena Coding · Top 3%#3 Q3 Can it find better answers to optimization problems with no single right answer? ALE-Bench · Top 13%#11 Q1 Top 6%#5 of 84 57.1% Q2 1552 Q3 2100 Q4 1682 Q5 61.8% Q6 85.7 Q7 47.6% Q8 77.9% Q9 1547 Q10 2.5% Q11 91.9% Q12 56.5% Q13 55.0% Q14 58.9% Q15 78.7% Q16 40.4% 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 · Top 2%#2 Q1 Which AI gives the writing and literature answers people prefer most? Arena Writing, Literature & Language · Top 2%#2 Q2 Which AI do people prefer most in long conversations? Arena Multi-turn · Top 2%#2 Q3 Which AI writes the creative pieces people prefer most? Arena Creative Writing · Top 2%#2 Q1 Top 2%#2 of 108 1521 Q2 1542 Q3 1520 Q4 1526 Q5 1236 Q6 87.4% Q7 1513 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 3%#3 Q1 Can it solve problems that stump experts across fields? Humanity's Last Exam · Top 3%#3 Q2 Can it solve research-level physics problems? CritPt · Top 18%#15 Q3 Which AI gives the math answers people prefer most? Arena Math · Top 1%#1 Q1 Top 3%#3 of 122 57.1% Q2 27.1% Q3 1535 Q4 100.0% Q5 86.0% Q6 78.0% Q7 99.0% Q8 1540 Q9 1542 Q10 95.1 Q11 1.5% Q12 98.5% Q13 91.3% Q14 91.7% Q15 59.0% Q16 64.0% Q17 97.6% Q18 63.7 About Humanity's Last Exam More about Humanity's Last Exam
Q1 Can it spot the wrong step from assembly photos? Furniture Assembly · Top 27%#10 Q1 Can it spot the wrong step from assembly photos? Furniture Assembly · Top 27%#10 Q2 Can it draw a floor plan from a few photos of a home? Blueprint-Bench 2 · Top 14%#5 Q3 Can it read charts and answer accurately? Chartography · Top 17%#1
Q1 Top 27%#10 of 38 72.5% Reasoning effort: High Q2 54.4% Q3 71.6% About Furniture Assembly More about Furniture Assembly
Q1 Which AI do people prefer most when asking in Korean? Arena Korean · Top 4%#3 Q1 Which AI do people prefer most when asking in Korean? Arena Korean · Top 4%#3 Q2 Which AI do people prefer most when asking in Japanese? Arena Japanese · Top 4%#3 Q3 Which AI do people prefer most on tough questions? Arena Hard Prompts · Top 2%#2 Q1 Top 4%#3 of 91 1487 Q2 1522 Q3 1544 Q4 1549 Q5 1534 Q6 1527 Q7 64.8% Q8 78.6% Q9 73.4% Q10 54.2% Q11 71.5% Q12 72.2% Q13 84.0% Q14 89.9% Q15 72.4% Q16 95.4% Q17 96.3% Q18 94.5% Q19 1371 Q20 916 Q21 536 Q22 75.6% About Arena Korean More about Arena Korean
Q1 Does it call out a nonsense question instead of playing along? BullshitBench v2 · Top 20%#28 Q1 Does it call out a nonsense question instead of playing along? BullshitBench v2 · Top 20%#28 Q2 Can it sort word-puzzle tiles into groups using meaning and wordplay? NYT Connections Extended · Top 3%#3 Q3 What would it score on Korea's 2026 CSAT (Suneung)? CSAT 2026 · Top 7%#4 Q1 Top 20%#28 of 143 66.0% Reasoning effort: Low Q2 98.6% Q3 446 Q4 450 Q5 39.3 Q6 1686 Q7 1083 Q8 1121 By effort Low 66.0% High 65.0%
About BullshitBench v2 More about BullshitBench v2