Leaderboard
AI model leaderboard
Which models people prefer when they compare two answers side by side, from LMArena’s open votes. It measures preference, not correctness, and the ± range shows how sure the ranking is. Prices from our API price table.
Building web apps from a prompt. Leaderboard of 2026-11-09. Checked .
| # | Model | Score | Votes | Licence | API $/M in · out |
|---|---|---|---|---|---|
| 1 | Claude Opus 5.5 (Max) anthropic | 1818 ±12 | 4,009 | Proprietary | — |
| 2 | Claude Sonnet 5.5 (xHigh) anthropic | 1774 ±10 | 5,493 | Proprietary | — |
| 3 | Gemini 4 Carbon (High) google | 1768 ±15 | 2,128 | Proprietary | — |
| 4 | GPT-6 Astra (Max) openai | 1767 ±9 | 8,117 | Proprietary | — |
| 5 | Claude Fable 5.1 (Max) anthropic | 1745 ±8 | 9,453 | Proprietary | — |
| 6 | GPT-6.1 Sol (Max) openai | 1743 ±10 | 5,250 | Proprietary | — |
| 7 | Muse Wonder Full (Max) meta | 1724 ±14 | 2,080 | Proprietary | — |
| 8 | Claude Sonnet 5.5 (High) anthropic | 1713 ±9 | 7,541 | Proprietary | — |
| 9 | Claude Opus 5 (Max) anthropic | 1691 ±6 | 19,593 | Proprietary | — |
| 10 | Kimi K3.1 Pro (Max) moonshot | 1688 ±10 | 4,591 | Kimi K3.1 Pro License | — |
| 11 | Gemini 4 Argon (High) google | 1683 ±10 | 4,626 | Proprietary | — |
| 12 | GPT-6.1 Terra (Max) openai | 1679 ±11 | 3,559 | Proprietary | — |
| 13 | Qwen3.8-Max (0902) alibaba | 1677 ±6 | 21,354 | Proprietary | — |
| 14 | GPT-6 Sol (Max) openai | 1674 ±9 | 6,882 | Proprietary | — |
| 15 | Qwen3.8-Max alibaba | 1670 ±12 | 3,373 | Proprietary | $2 · $6 |
| 16 | DeepSeek-V4.1-Flash-20261017 (Max) deepseek | 1664 ±7 | 13,241 | MIT | — |
| 17 | Kimi K3 (Max) moonshot | 1657 ±6 | 22,114 | Kimi K3 license | — |
| 18 | Claude Opus 5 (High) anthropic | 1657 ±5 | 26,982 | Proprietary | — |
| 19 | Muse Spark 1.3 (Max) meta | 1654 ±7 | 11,758 | Proprietary | — |
| 20 | Muse Wonder (Max) meta | 1650 ±10 | 4,875 | Proprietary | — |
| 21 | Inkling 2 Small thinky | 1649 ±20 | 964 | Proprietary | — |
| 22 | Grok 4.7 (xHigh) xai | 1639 ±8 | 7,816 | Proprietary | — |
| 23 | MiMo-V2.6-Flash xiaomi | 1637 ±8 | 8,576 | MIT | — |
| 24 | Hy4 preview tencent | 1631 ±10 | 4,606 | Apache 2.0 | — |
| 25 | Claude Fable 5 (High) anthropic | 1625 ±6 | 17,017 | Proprietary | — |
| 26 | Qwen3.8-Flash-Next alibaba | 1624 ±7 | 10,939 | qwen-community-1.0 | — |
| 27 | MiMo-V2.6-Pro xiaomi | 1624 ±8 | 8,402 | MIT | — |
| 28 | GLM-5.3 (Max) zai | 1622 ±7 | 12,263 | MIT | — |
| 29 | DeepSeek-V4.1-Flash-20260910 (Max) deepseek | 1621 ±11 | 3,408 | MIT | — |
| 30 | Muse Spark 1.3 (xHigh) meta | 1621 ±7 | 11,277 | Proprietary | — |
| 31 | Grok 4.6 (High) xai | 1618 ±7 | 12,193 | Proprietary | — |
| 32 | GPT-5.6 Sol (xHigh, Codex Harness) openai | 1615 ±6 | 18,929 | Proprietary | — |
| 33 | GLM-5.3-Flash zai | 1609 ±6 | 23,185 | MIT | $0.15 · $0.5 |
| 34 | Muse Ember (Max) meta | 1608 ±17 | 1,345 | Proprietary | — |
| 35 | GLM-5.2 (Max) zai | 1608 ±6 | 17,934 | MIT | — |
| 36 | Claude Haiku 5.5 (High) anthropic | 1595 ±9 | 6,174 | Proprietary | — |
| 37 | Qwen3.8-27B alibaba | 1591 ±6 | 20,408 | Apache 2.0 | — |
| 38 | Gemini 3.7 Flash (High) google | 1588 ±6 | 14,619 | Proprietary | — |
| 39 | Gemini 3.8 Flash (High) google | 1584 ±7 | 13,485 | Proprietary | — |
| 40 | DeepSeek-V4-Flash (High) deepseek | 1582 ±9 | 4,839 | MIT | — |
| 41 | MiniMax-M3.1-Flash minimax | 1582 ±10 | 4,421 | MiniMax Community License | — |
| 42 | DeepSeek-V4-Pro (High, 20260813) deepseek | 1581 ±10 | 4,614 | MIT | — |
| 43 | GPT-6 Luna (Max) openai | 1578 ±6 | 29,989 | Proprietary | — |
| 44 | Step 5 Preview (High) stepfun | 1572 ±8 | 7,267 | Proprietary | — |
| 45 | Mistral Large 4 mistral | 1562 ±17 | 1,409 | Modified MIT | $0.7 · $2.1 |
| 46 | Claude Opus 4.7 (High) anthropic | 1557 ±6 | 17,616 | Proprietary | — |
| 47 | Claude Opus 4.8 (High) anthropic | 1555 ±6 | 18,061 | Proprietary | — |
| 48 | Claude Opus 4.7 anthropic | 1555 ±6 | 17,957 | Proprietary | $5 · $25 |
| 49 | Grok 4.5 xai | 1552 ±7 | 10,815 | Proprietary | $2 · $6 |
| 50 | Claude Opus 4.6 (High) anthropic | 1546 ±6 | 19,080 | Proprietary | — |
| 51 | Muse Spark 1.1 meta | 1542 ±8 | 7,850 | Proprietary | — |
| 52 | Claude Sonnet 5 (High) anthropic | 1539 ±6 | 16,802 | Proprietary | — |
| 53 | Gemini 3.6 Flash (High) google | 1538 ±8 | 8,538 | Proprietary | — |
| 54 | Claude Opus 4.6 anthropic | 1537 ±5 | 20,385 | Proprietary | $5 · $25 |
| 55 | Claude Opus 4.8 anthropic | 1536 ±6 | 16,993 | Proprietary | $5 · $25 |
| 56 | Muse Spark 1.2 (xHigh) meta | 1534 ±14 | 2,121 | Proprietary | — |
| 57 | Muse Spark Open (Max) meta | 1526 ±15 | 1,780 | muse-model-license | — |
| 58 | Claude Sonnet 4.6 anthropic | 1522 ±5 | 22,558 | Proprietary | $3 · $15 |
| 59 | GPT-5.6 Terra (xHigh, Codex Harness) openai | 1520 ±6 | 15,502 | Proprietary | — |
| 60 | Seed2.1 Pro Preview bytedance | 1519 ±6 | 17,367 | Proprietary | — |
| 61 | GPT-5.6 Luna (xHigh, Codex Harness) openai | 1519 ±6 | 12,231 | Proprietary | — |
| 62 | Qwen3.7-Max (20260517) alibaba | 1515 ±7 | 8,751 | Proprietary | — |
| 63 | GPT-5.5 (xHigh, Codex Harness) openai | 1512 ±6 | 15,398 | Proprietary | — |
| 64 | Kimi K2.6 moonshot | 1509 ±7 | 10,016 | Modified MIT | $0.95 · $4 |
| 65 | GLM-5.1 zai | 1508 ±7 | 10,543 | MIT | $1.4 · $4.4 |
| 66 | Hy3 tencent | 1506 ±8 | 7,207 | Apache 2.0 | — |
| 67 | Gemini 3.5 Flash (High) google | 1500 ±7 | 9,167 | Proprietary | — |
| 68 | Claude Opus 4.5 (High, 32K, 20251101) anthropic | 1493 ±8 | 10,930 | Proprietary | — |
| 69 | Gemini 3.5 Flash (Medium) google | 1490 ±7 | 11,913 | Proprietary | — |
| 70 | GPT-5.5 (High, Codex Harness) openai | 1487 ±6 | 17,673 | Proprietary | — |
| 71 | Spark-X2.6 synlan | 1485 ±19 | 1,189 | Proprietary | — |
| 72 | MiniMax-M3 minimax | 1484 ±5 | 21,077 | MiniMax Community License | — |
| 73 | Qwen3.6-Max-Preview alibaba | 1482 ±13 | 2,677 | Proprietary | — |
| 74 | MiMo-V2.5-Pro xiaomi | 1479 ±5 | 22,101 | MIT | — |
| 75 | Kimi K2.7 Code moonshot | 1474 ±9 | 4,891 | Modified MIT | $0.95 · $4 |
| 76 | Claude Opus 4.5 (20251101) anthropic | 1469 ±7 | 13,310 | Proprietary | — |
| 77 | DeepSeek-V4-Pro Preview (High) deepseek | 1463 ±6 | 12,958 | MIT | — |
| 78 | GPT-5.4 (High, Codex Harness) openai | 1462 ±19 | 1,340 | Proprietary | — |
| 79 | Qwen3.6-Plus alibaba | 1461 ±6 | 18,273 | Proprietary | — |
| 80 | GPT-5.5 (Codex Harness) openai | 1455 ±6 | 15,634 | Proprietary | — |
| 81 | DeepSeek-V4-Pro deepseek | 1448 ±6 | 13,779 | MIT | $1.32 · $3.96 |
| 82 | Gemini 3.1 Pro Preview google | 1446 ±5 | 23,865 | Proprietary | $2 · $12 |
| 83 | GPT-5.4 (Medium, Codex Harness) openai | 1442 ±18 | 1,363 | Proprietary | — |
| 84 | Gemini 3 Pro google | 1441 ±8 | 13,937 | Proprietary | — |
| 85 | Gemini 3 Flash google | 1440 ±9 | 10,793 | Proprietary | $0.5 · $3 |
| 86 | Beam 1.0 501B A23B reflection | 1438 ±14 | 2,521 | Apache 2.0 | — |
| 87 | MiMo-V2.5 xiaomi | 1438 ±6 | 14,646 | MIT | — |
| 88 | Kimi K2.5 (Thinking) moonshot | 1436 ±5 | 19,815 | Modified MIT | — |
| 89 | Gemini 3.5 Flash-Lite google | 1435 ±41 | 233 | Proprietary | $0.3 · $2.5 |
| 90 | GLM-5 zai | 1435 ±8 | 7,418 | MIT | $1 · $3.2 |
| 91 | GLM-4.7 zai | 1435 ±12 | 3,975 | MIT | $0.6 · $2.2 |
| 92 | MiMo-V2-Pro xiaomi | 1434 ±8 | 7,194 | Proprietary | — |
| 93 | DeepSeek-V4-Flash Preview (High) deepseek | 1431 ±7 | 9,986 | MIT | — |
| 94 | GPT-5.2 openai | 1420 ±23 | 1,065 | Proprietary | $1.75 · $14 |
| 95 | GPT-5 (Medium) openai | 1417 ±16 | 3,036 | Proprietary | — |
| 96 | Inkling thinky | 1414 ±6 | 17,206 | Apache 2.0 | — |
| 97 | Inkling-Small thinky | 1407 ±9 | 5,228 | Apache 2.0 | — |
| 98 | GPT-5.3-Codex (Codex Harness) openai | 1406 ±14 | 2,630 | Proprietary | — |
| 99 | Kimi K2.5 (Instant) moonshot | 1406 ±12 | 3,244 | Modified MIT | — |
| 100 | Qwen3.5-397B-A17B alibaba | 1400 ±5 | 20,871 | Apache 2.0 | — |
Data: LMArena leaderboard dataset, licensed CC BY 4.0. Prices are matched by model name and may be missing for models without a public API.