AI Rankings
Human Preference Leaderboard: AI Models Ranked by Blind Votes
On Arena, people send the same question to two anonymous AI models and vote for the better answer. This page ranks models by the rating those votes produce. It shows what people like to read, which is not always the same as which model is most accurate.
Data updated
Top 20 at a glance
Tap a model to see its full page: scores, price and how it compares.
Bars start at 1470, not zero, so differences are easier to see.
Full ranking
Showing the top 100 of 166 models.Download the full table (CSV)Data credit: Arena (formerly LMArena), Epoch AI, LiteLLM model price list, used under open licences (details at the foot of the page).
| Rank and model | Arena text rating | Likely range | Votes | Epoch index | Price per 1M tokens (in / out) |
|---|---|---|---|---|---|
1Gemini 4 ArgonGoogle | 1517 | 1511 to 1524 | 10,083 | 166.4 | $2.00 / $10.00 |
2Claude Opus 5.5Anthropic | 1509 | 1502 to 1516 | 10,322 | 168.8 | $4.00 / $20.00 |
2Muse WonderMeta | 1509 | 1501 to 1517 | 6,001 | 161.8 | Not measured |
4Claude Fable 5Anthropic | 1503 | 1499 to 1508 | 44,253 | 163.4 | $10.00 / $50.00 |
5Claude Fable 5.1Anthropic | 1503 | 1497 to 1509 | 16,343 | 166.5 | $10.00 / $50.00 |
6Claude Opus 4.6Anthropic | 1503 | 1500 to 1506 | 83,454 | 155.7 | $5.00 / $25.00 |
7Claude Opus 4.7Anthropic | 1501 | 1497 to 1505 | 65,140 | 156.8 | $5.00 / $25.00 |
8Gemini 3.8 FlashGoogle | 1496 | 1491 to 1500 | 43,997 | 157.4 | $0.75 / $3.75 |
9Muse Spark 1.3Meta | 1496 | 1490 to 1501 | 20,506 | 157.2 | $1.25 / $4.25 |
10Muse Spark 1.2Meta | 1491 | 1485 to 1498 | 10,470 | 155.1 | $1.25 / $4.25 |
11Claude Opus 5Anthropic | 1491 | 1486 to 1496 | 34,526 | 163.9 | $5.00 / $25.00 |
12Muse Spark 1.1Meta | 1490 | 1486 to 1494 | 46,121 | 154.1 | $1.25 / $4.25 |
13Gemini 3.7 FlashGoogle | 1490 | 1485 to 1495 | 30,501 | 158.0 | $0.75 / $3.75 |
14Muse SparkMeta | 1489 | 1484 to 1495 | 14,147 | 152.0 | Not measured |
15Kimi K3Moonshot AI · Open weights | 1487 | 1483 to 1492 | 35,379 | 157.9 | $3.00 / $15.00 |
16Gemini 3.1 ProGoogle | 1487 | 1484 to 1490 | 130,819 | 155.3 | $2.00 / $12.00 |
17Gemini 3 ProGoogle | 1486 | 1482 to 1489 | 42,054 | 153.1 | $2.00 / $12.00 |
18GPT-5.6 SolOpenAI | 1485 | 1481 to 1489 | 42,713 | 162.8 | $4.00 / $20.00 |
19Gemini 3.6 FlashGoogle | 1483 | 1479 to 1488 | 36,102 | 154.5 | $0.75 / $3.75 |
20Qwen 3.8 MaxAlibaba | 1483 | 1478 to 1488 | 30,500 | 157.4 | $2.00 / $6.00 |
21GPT-5.5OpenAI | 1482 | 1478 to 1485 | 69,375 | 159.7 | $5.00 / $30.00 |
22Claude Opus 4.8Anthropic | 1482 | 1478 to 1485 | 64,096 | 159.0 | $5.00 / $25.00 |
23GLM-5.3Z.ai · Open weights | 1481 | 1476 to 1486 | 25,144 | 156.0 | $1.40 / $4.40 |
24GPT-6.1 SolOpenAI | 1480 | 1473 to 1487 | 8,305 | 168.4 | $2.00 / $10.00 |
25MiMo-V2.6-ProXiaomi · Open weights | 1480 | 1470 to 1489 | 4,067 | Not measured | $0.43 / $0.87 |
26Claude Sonnet 5.5Anthropic | 1478 | 1471 to 1485 | 8,977 | 166.2 | $2.00 / $10.00 |
27Gemini 3.5 FlashGoogle | 1477 | 1473 to 1481 | 48,504 | 154.9 | $1.50 / $9.00 |
28GPT-6 AstraOpenAI | 1477 | 1471 to 1483 | 14,076 | 168.8 | $10.00 / $50.00 |
29DeepSeek V4.1 FlashDeepSeek · Open weights | 1476 | 1470 to 1482 | 13,750 | 155.9 | $0.30 / $1.20 |
29GLM-5.3-FlashZ.ai · Open weights | 1476 | 1471 to 1481 | 38,359 | 153.1 | $0.15 / $0.50 |
31GPT-5.2 ChatOpenAI | 1476 | 1472 to 1480 | 35,982 | Not measured | $1.75 / $14.00 |
32GLM-5.2Z.ai · Open weights | 1475 | 1471 to 1480 | 52,283 | 151.9 | $1.40 / $4.40 |
32GPT-5.4OpenAI | 1475 | 1472 to 1479 | 64,757 | 157.3 | $2.50 / $15.00 |
34Grok 4.20 Beta 1xAI | 1475 | 1470 to 1479 | 27,914 | Not measured | Not measured |
35Qwen3.7-MaxAlibaba | 1474 | 1464 to 1484 | 3,941 | 154.0 | $2.50 / $7.50 |
36Claude Opus 4.5Anthropic | 1474 | 1470 to 1477 | 37,485 | 150.1 | $5.00 / $25.00 |
37GPT-5.5 InstantOpenAI | 1473 | 1468 to 1478 | 27,163 | 142.4 | Not measured |
38Gemini 3 FlashGoogle | 1473 | 1469 to 1477 | 31,313 | 151.9 | $0.50 / $3.00 |
39Claude Sonnet 4.6Anthropic | 1472 | 1469 to 1476 | 70,953 | 152.8 | $3.00 / $15.00 |
| 1472 | 1468 to 1475 | 66,418 | Not measured | $1.25 / $2.50 | |
| 1471 | 1467 to 1474 | 64,928 | Not measured | $1.25 / $2.50 | |
42GPT-5.6 TerraOpenAI | 1467 | 1463 to 1471 | 46,470 | 160.1 | $2.00 / $12.00 |
43ERNIE 5.1Baidu | 1467 | 1463 to 1472 | 39,474 | Not measured | $0.56 / $2.54 |
44MiMo-V2.5-ProXiaomi · Open weights | 1467 | 1463 to 1470 | 77,883 | Not measured | $0.52 / $1.04 |
45Grok 4.5xAI | 1466 | 1461 to 1470 | 39,880 | 154.1 | $2.00 / $6.00 |
46GLM-5.1Z.ai · Open weights | 1465 | 1462 to 1469 | 65,402 | 149.8 | $1.40 / $4.40 |
47Grok 4.1xAI | 1465 | 1462 to 1468 | 66,479 | Not measured | Not measured |
48Qwen3.5-Max-PreviewAlibaba | 1465 | 1460 to 1469 | 22,601 | Not measured | Not measured |
49DeepSeek V4 Pro 0813DeepSeek · Open weights | 1465 | 1458 to 1471 | 9,797 | 155.6 | $1.32 / $3.96 |
50Claude Sonnet 5Anthropic | 1462 | 1457 to 1466 | 51,784 | 156.9 | $2.00 / $10.00 |
51Kimi K2.6Moonshot AI · Open weights | 1461 | 1456 to 1465 | 40,026 | 150.9 | $0.95 / $4.00 |
52Qwen 3.6 Max (Preview)Alibaba | 1461 | 1452 to 1469 | 5,423 | 149.4 | $1.27 / $7.61 |
53GPT-6 SolOpenAI | 1459 | 1453 to 1465 | 14,707 | 163.7 | $2.00 / $10.00 |
54DeepSeek-V4-ProDeepSeek · Open weights | 1458 | 1454 to 1462 | 57,588 | 149.1 | $1.32 / $3.96 |
55GLM-5Z.ai · Open weights | 1457 | 1453 to 1462 | 29,267 | 145.8 | $1.00 / $3.20 |
56Grok 4.6xAI | 1456 | 1452 to 1461 | 32,189 | 156.6 | $2.00 / $6.00 |
57Claude Sonnet 4.5Anthropic | 1456 | 1454 to 1459 | 83,906 | 146.7 | $3.00 / $15.00 |
58GPT-5.1OpenAI | 1456 | 1452 to 1460 | 41,646 | 149.8 | $1.25 / $10.00 |
58Qwen3.7-PlusAlibaba | 1456 | 1451 to 1460 | 42,058 | 147.0 | $0.36 / $1.44 |
60Gemini 3.5 Flash-LiteGoogle | 1455 | 1451 to 1460 | 35,829 | 144.9 | $0.30 / $2.50 |
61GPT-5.6 LunaOpenAI | 1453 | 1449 to 1457 | 41,426 | 156.5 | $0.20 / $1.20 |
62Kimi K2.5Moonshot AI · Open weights | 1450 | 1447 to 1453 | 74,744 | 148.1 | $0.60 / $3.00 |
63Claude Opus 4.1Anthropic | 1450 | 1446 to 1453 | 49,536 | 144.1 | $15.00 / $75.00 |
64GPT-5.4 MiniOpenAI | 1447 | 1443 to 1451 | 63,451 | 148.7 | $0.75 / $4.50 |
65Gemini 2.5 Pro (Jun 2025)Google | 1446 | 1443 to 1448 | 125,316 | 145.1 | $1.25 / $10.00 |
66GPT-4.5OpenAI | 1445 | 1439 to 1450 | 14,547 | 136.7 | $75.00 / $150 |
67Qwen 3.6 PlusAlibaba | 1443 | 1439 to 1447 | 48,142 | 147.3 | $0.50 / $3.00 |
68GPT-6 LunaOpenAI | 1442 | 1438 to 1447 | 72,504 | 156.8 | $0.10 / $0.50 |
69Qwen3.5 397B-A17BAlibaba · Open weights | 1442 | 1439 to 1445 | 96,206 | 146.8 | $0.60 / $3.60 |
70InklingThinking Machines · Open weights | 1442 | 1437 to 1446 | 43,705 | 148.4 | $1.00 / $4.05 |
71GLM-4.7Z.ai · Open weights | 1441 | 1435 to 1447 | 12,245 | 143.6 | $0.60 / $2.20 |
72Grok 4.7xAI | 1441 | 1435 to 1447 | 14,693 | 154.4 | $2.00 / $6.00 |
73MiniMax-M3MiniMax · Open weights | 1440 | 1436 to 1444 | 67,030 | 146.9 | $0.30 / $1.20 |
74DeepSeek-V4-FlashDeepSeek · Open weights | 1439 | 1435 to 1443 | 51,781 | 146.1 | $0.30 / $1.20 |
75Gemma 4 26B A4BGoogle · Open weights | 1438 | 1430 to 1445 | 6,092 | 141.9 | $0.15 / $0.60 |
76GPT-5.2OpenAI | 1437 | 1434 to 1441 | 49,643 | 153.7 | $1.75 / $14.00 |
77Qwen 3.8 27BAlibaba · Open weights | 1437 | 1432 to 1442 | 26,238 | 151.7 | $0.45 / $3.00 |
78GPT-5OpenAI | 1435 | 1430 to 1439 | 31,712 | 150.0 | $1.25 / $10.00 |
79Qwen3-MaxAlibaba | 1434 | 1430 to 1439 | 27,367 | 142.3 | $1.66 / $7.22 |
80Mistral Large 4Mistral AI | 1434 | 1427 to 1440 | 10,331 | 150.0 | $0.68 / $2.09 |
81Gemini 3.1 Flash-LiteGoogle | 1433 | 1429 to 1437 | 64,209 | 144.3 | $0.25 / $1.50 |
82o3OpenAI | 1432 | 1428 to 1435 | 58,894 | 146.7 | $2.00 / $8.00 |
83Mistral Medium 3.5Mistral AI · Open weights | 1426 | 1420 to 1433 | 11,791 | 141.4 | $1.50 / $7.50 |
84Claude Opus 4Anthropic | 1426 | 1422 to 1431 | 35,908 | 142.5 | $15.00 / $75.00 |
85DeepSeek-V3.2-ExpDeepSeek · Open weights | 1425 | 1419 to 1432 | 8,994 | 144.9 | $0.27 / $0.41 |
86DeepSeek-V3.2DeepSeek · Open weights | 1425 | 1421 to 1428 | 48,090 | 146.1 | $0.28 / $0.40 |
87Claude Haiku 5.5Anthropic | 1425 | 1416 to 1433 | 6,080 | 159.3 | $0.10 / $0.50 |
88Qwen3-235B-A22B-Instruct (Jul 2025)Alibaba · Open weights | 1422 | 1420 to 1425 | 98,096 | 138.8 | $0.22 / $0.88 |
89DeepSeek-R1 (May 2025)DeepSeek · Open weights | 1421 | 1416 to 1427 | 18,170 | 141.1 | $0.75 / $2.45 |
90Grok 4 FastxAI | 1419 | 1411 to 1426 | 6,337 | 144.1 | Not measured |
91Kimi K2 (Jul 2025)Moonshot AI · Open weights | 1418 | 1413 to 1423 | 27,136 | 140.0 | $0.55 / $2.20 |
92DeepSeek-V3.1DeepSeek · Open weights | 1417 | 1411 to 1423 | 14,607 | 139.8 | $0.56 / $1.68 |
93MiniMax-M2.7MiniMax · Open weights | 1415 | 1412 to 1418 | 87,731 | 145.8 | $0.30 / $1.19 |
94GPT-4.1OpenAI | 1415 | 1411 to 1418 | 50,302 | 136.7 | $2.00 / $8.00 |
95Claude Haiku 4.5Anthropic | 1415 | 1412 to 1417 | 152,393 | 142.2 | $1.00 / $5.00 |
96Grok 3xAI | 1411 | 1407 to 1416 | 32,498 | 138.2 | $3.00 / $15.00 |
97Grok 4xAI | 1411 | 1407 to 1415 | 39,679 | 146.2 | $3.00 / $15.00 |
98Gemini 2.5 Flash (Sep 2025)Google | 1409 | 1407 to 1412 | 125,462 | 142.9 | $0.30 / $2.50 |
99Inkling-SmallThinking Machines · Open weights | 1405 | 1400 to 1409 | 35,359 | 150.3 | $0.47 / $1.20 |
100o1OpenAI | 1402 | 1398 to 1407 | 27,807 | 141.7 | $15.00 / $60.00 |
How to read this leaderboard
Higher is better. The rating goes up when a model wins more votes against strong opponents, so only the gaps between models matter. This board uses Arena's style-controlled ratings, which adjust for answer length and formatting so that longer or more decorated answers do not win by default. Each rating has a likely range: overlapping ranges mean the models are effectively tied.
Questions people ask
How do Arena ratings work?
People compare two anonymous answers and pick the better one. A rating system turns those head-to-head results into a score for each model. A model that keeps beating strong models rises, and one that loses falls.
Is a popular answer the same as a correct answer?
No. Votes capture what people prefer, including tone and clarity. For tasks where being exactly right matters, check benchmark results as well, such as the Epoch Capabilities Index on the overall leaderboard.
Does the top model suit my work best?
Not always. A leaderboard compares models on one kind of test. Price, speed, privacy rules, the apps you already use and the type of work you do all matter. Use the rank as a starting point, then try two or three models on your own tasks.
Why is a model I know missing?
A model appears only when an open source has published results for it. New models can take days or weeks to be measured, and models with no published score are left out rather than guessed.
Where does the data come from?
Ratings come from the open leaderboard dataset published by Arena (arena.ai), licensed CC BY 4.0. We keep the latest published ratings, match models to our other sources and rank them. Arena does not endorse this site.
How often is this leaderboard updated?
We check our open data sources every day and show the date of the data at the top of the page. A model only moves when its source publishes new results, so a number can stay the same for weeks.
More leaderboards
Different questions need different tests. Pick the one that matches your work.
- Overall intelligenceWhich AI model is the smartest overall?Leads: Claude Opus 5.5 (168.8)270 models ranked
- CodingWhich AI model is best for coding?Leads: Gemini 4 Argon (1552)166 models ranked
- Web developmentWhich AI model builds the best websites and apps?Leads: Claude Opus 5.5 (1807)85 models ranked
- Image understandingWhich AI model understands images best?Leads: Claude Fable 5 (1309)85 models ranked
- Open-weights modelsWhich open-source AI model is best?Leads: Kimi K3 (157.9)123 models ranked
- Cheapest strong modelsWhat is the cheapest AI model that is still good?Cheapest: gpt-oss-20b ($0.10)119 models ranked
- Best valueWhich AI models give the most intelligence for the money?Line starts at: gpt-oss-20b ($0.10)7 models ranked
- Context windowWhich AI model can read the most text at once?Leads: GPT-5.4 (1.1M tokens)193 models ranked
- GPQA DiamondWhich AI model scores highest on GPQA Diamond?Leads: Gemini 4 Argon (96.4%)183 models ranked
- Humanity's Last ExamWhich AI model scores highest on Humanity's Last Exam?Leads: GPT-6 Astra (54.8%)41 models ranked
- SWE-bench VerifiedWhich AI model scores highest on SWE-bench Verified?Leads: Claude Opus 4.7 (83.5%)32 models ranked
- Aider PolyglotWhich AI model scores highest on Aider Polyglot?Leads: GPT-5 (88.0%)46 models ranked
- Terminal-BenchWhich AI model scores highest on Terminal-Bench?Leads: GPT-5.3 Codex (64.7%)25 models ranked
- FrontierMath (Tiers 1-3)Which AI model scores highest on FrontierMath (Tiers 1-3)?Leads: GPT-6 Astra (93.7%)87 models ranked
- ARC-AGI-2Which AI model scores highest on ARC-AGI-2?Leads: GPT-6 Astra (95.0%)86 models ranked
- SimpleQA VerifiedWhich AI model scores highest on SimpleQA Verified?Leads: Gemini 4 Argon (76.0%)82 models ranked
- SimpleBenchWhich AI model scores highest on SimpleBench?Leads: Claude Fable 5 (81.9%)80 models ranked
- DeepSWEWhich AI model scores highest on DeepSWE?Leads: GPT-6 Astra (74.1%)26 models ranked
Where this data comes from
- Arena (formerly LMArena): Human-preference ratings: Arena (arena.ai) leaderboard dataset, CC BY 4.0. We keep the latest published ratings, match them to models from other sources and rank them. Ratings are not changed. Arena does not endorse this site.
- Epoch AI: Benchmark data: Epoch AI, 'Capabilities & benchmarking', epoch.ai/benchmarks (CC BY 4.0). We match models across sources, rank them and show them in charts. Scores are not changed. Epoch AI does not endorse this site.
- LiteLLM model price list: Prices and context windows: LiteLLM model price list (github.com/BerriAI/litellm, MIT licence). We use the model maker's own price entry when there is one. Prices change often, so always confirm on the maker's site before you buy.
Newest data on this page: 26 Oct 2026. Prices are list prices and change often, so confirm on the maker's site before you buy. Sponsors never change a score or an order. How we calculate this · Report a mistake