Benchmark · Arenas: people's votes

LMArena Code Arena (WebDev)

Which of two anonymous models' work people prefer on web development prompts in the arena's Code Arena, fitted as a Bradley-Terry rating.

Score
arena rating (Bradley-Terry), higher is better
Run by
LMArena
Status
Current, newest result 19 November 2026
Updated
daily
Data
the published file, CC-BY-4.0, read 20 November 2026
Method
arena.ai/blog/leaderboard-changelog/

Results

53 models we list, each at its best setting, best first.

#ModelScoreRangeWho ran itSettingDate
1Claude Fable 5.2 A18231807 to 1839Independenteffort xhigh2026-11-19source
2Claude Opus 5.5 A18181807 to 1829Independenteffort max2026-11-19source
3Claude Sonnet 5.5 A17711762 to 1781Independenteffort xhigh2026-11-19source
4Gemini 4 Carbon B17691753 to 1784Independenteffort high2026-11-19source
5Gemini 4 Neon B17631750 to 1777Independenteffort high2026-11-19source
6GPT-6 Astra A17631755 to 1771Independenteffort max2026-11-19source
7GPT-6.2 Astra AA17591744 to 1775Independenteffort max2026-11-19source
8GPT-6.2 Sol AA17541740 to 1770Independenteffort max2026-11-19source
9GPT-6.1 Sol AA17441734 to 1753Independenteffort max2026-11-19source
10Claude Fable 5.1 A17411734 to 1749Independenteffort max2026-11-19source
11Gemini 4 Hydrogen B17011687 to 1715Independenteffort high2026-11-19source
12GPT-6.1 Terra AA16841675 to 1694Independenteffort max2026-11-19source
13Kimi K3.1 Pro D16831674 to 1692Independenteffort max2026-11-19source
14Gemini 4 Argon B16811672 to 1690Independenteffort high2026-11-19source
15GPT-6 Sol A16741666 to 1682Independenteffort max2026-11-19source
16DeepSeek V4.1 Flash C16591652 to 1665Independenteffort max2026-11-19source
17Kimi K3 B16581653 to 1664Independenteffort max2026-11-19source
18Muse Spark 1.3 D16541647 to 1660Independenteffort max2026-11-19source
19Tencent Hy4 B16501638 to 1661Independent2026-11-19source
20Muse Wonder D16481639 to 1656Independenteffort max2026-11-19source
21Inkling 2 Small C16441626 to 1661Independent2026-11-19source
22Grok 4.7 BB16381630 to 1646Independenteffort xhigh2026-11-19source
23Qwen3.8-Flash-Next C16261619 to 1633Independent2026-11-19source
24GLM-5.3 B16221615 to 1628Independenteffort max2026-11-19source
25GLM-5.3-Flash B16101604 to 1616Independent2026-11-19source
26Muse Ember D16091592 to 1626Independenteffort max2026-11-19source
27Claude Haiku 5.5 A15961588 to 1604Independenteffort high2026-11-19source
28Qwen3.8-27B B15891583 to 1595Independent2026-11-19source
29Gemini 3.8 Flash AA15841578 to 1590Independenteffort high2026-11-19source
30MiniMax-M3.1-Flash C15821573 to 1590Independent2026-11-19source
31DeepSeek V4 Pro D15811571 to 1590Independenteffort high, snapshot 202608132026-11-19source
32GPT-6 Luna A15791573 to 1585Independenteffort max2026-11-19source
33Step 5 Preview E15701562 to 1578Independenteffort high2026-11-19source
34Mistral Large 4 C15701557 to 1582Independent2026-11-19source
35Muse Spark Open C15271513 to 1541Independenteffort max2026-11-19source
36Hy3 B15061498 to 1513Independent2026-11-19source
37MiniMax M3 C14841479 to 1490Independent2026-11-19source
38Gemini 3.1 Pro (preview) AA14461441 to 1451Independent2026-11-19source
39Gemini 3.5 Flash-Lite AA14321390 to 1473Independent2026-11-19source
40Inkling B14141408 to 1420Independent2026-11-19source
41Inkling-Small B14081399 to 1417Independent2026-11-19source
42Nemotron 3.5 Super VL C14021392 to 1412Independentprecision NVFP42026-11-19source
43Qwen3.5-397B-A17B B14001395 to 1405Independent2026-11-19source
44MiniMax-M2.7 C13981392 to 1404Independent2026-11-19source
45MiniMax-M2.5 C13861378 to 1394Independent2026-11-19source
46Gemma 4 31B C13701364 to 1376Independent2026-11-19source
47DeepSeek-V3.2 C13621353 to 1372Independentthinking on2026-11-19source
48Qwen3.5-122B-A10B B13601352 to 1368Independent2026-11-19source
49Gemma 4 26B A4B IT BB13581341 to 1375Independent2026-11-19source
50Grok 4.3 BB13571351 to 1364Independent2026-11-19source
51Laguna M.1 C13501341 to 1359Independent2026-11-19source
52Mistral Medium 3.5 B12631248 to 1278Independent2026-11-19source
53Mistral Large 3 B12331207 to 1258Independent2026-11-19source
Every setting, 58 rows
ModelAs the leaderboard names itScoreWho ran itSettingDate
Claude Fable 5.2Claude Fable 5.2 (xHigh)1823Independenteffort xhigh2026-11-19source
Claude Opus 5.5Claude Opus 5.5 (Max)1818Independenteffort max2026-11-19source
Claude Sonnet 5.5Claude Sonnet 5.5 (xHigh)1771Independenteffort xhigh2026-11-19source
Gemini 4 CarbonGemini 4 Carbon (High)1769Independenteffort high2026-11-19source
Gemini 4 NeonGemini 4 Neon (High)1763Independenteffort high2026-11-19source
GPT-6 AstraGPT-6 Astra (Max)1763Independenteffort max2026-11-19source
GPT-6.2 AstraGPT-6.2 Astra (Max)1759Independenteffort max2026-11-19source
GPT-6.2 SolGPT-6.2 Sol (Max)1754Independenteffort max2026-11-19source
GPT-6.1 SolGPT-6.1 Sol (Max)1744Independenteffort max2026-11-19source
Claude Fable 5.1Claude Fable 5.1 (Max)1741Independenteffort max2026-11-19source
Claude Sonnet 5.5Claude Sonnet 5.5 (High)1716Independenteffort high2026-11-19source
Gemini 4 HydrogenGemini 4 Hydrogen (High)1701Independenteffort high2026-11-19source
GPT-6.1 TerraGPT-6.1 Terra (Max)1684Independenteffort max2026-11-19source
Kimi K3.1 ProKimi K3.1 Pro (Max)1683Independenteffort max2026-11-19source
Gemini 4 ArgonGemini 4 Argon (High)1681Independenteffort high2026-11-19source
GPT-6 SolGPT-6 Sol (Max)1674Independenteffort max2026-11-19source
DeepSeek V4.1 FlashDeepSeek-V4.1-Flash-20261017 (Max)1659Independenteffort max2026-11-19source
Kimi K3Kimi K3 (Max)1658Independenteffort max2026-11-19source
Muse Spark 1.3Muse Spark 1.3 (Max)1654Independenteffort max2026-11-19source
Tencent Hy4Hy41650Independent2026-11-19source
Muse WonderMuse Wonder (Max)1648Independenteffort max2026-11-19source
Inkling 2 SmallInkling 2 Small1644Independent2026-11-19source
Grok 4.7Grok 4.7 (xHigh)1638Independenteffort xhigh2026-11-19source
Qwen3.8-Flash-NextQwen3.8-Flash-Next1626Independent2026-11-19source
GLM-5.3GLM-5.3 (Max)1622Independenteffort max2026-11-19source
DeepSeek V4.1 FlashDeepSeek-V4.1-Flash-20260910 (Max)1621Independenteffort max2026-11-19source
Muse Spark 1.3Muse Spark 1.3 (xHigh)1621Independenteffort xhigh2026-11-19source
GLM-5.3-FlashGLM-5.3-Flash1610Independent2026-11-19source
Muse EmberMuse Ember (Max)1609Independenteffort max2026-11-19source
Claude Haiku 5.5Claude Haiku 5.5 (High)1596Independenteffort high2026-11-19source
Qwen3.8-27BQwen3.8-27B1589Independent2026-11-19source
Gemini 3.8 FlashGemini 3.8 Flash (High)1584Independenteffort high2026-11-19source
MiniMax-M3.1-FlashMiniMax-M3.1-Flash1582Independent2026-11-19source
DeepSeek V4 ProDeepSeek-V4-Pro (High, 20260813)1581Independenteffort high, snapshot 202608132026-11-19source
GPT-6 LunaGPT-6 Luna (Max)1579Independenteffort max2026-11-19source
Step 5 PreviewStep 5 Preview (High)1570Independenteffort high2026-11-19source
Mistral Large 4Mistral Large 41570Independent2026-11-19source
Muse Spark OpenMuse Spark Open (Max)1527Independenteffort max2026-11-19source
Hy3Hy31506Independent2026-11-19source
MiniMax M3MiniMax-M31484Independent2026-11-19source
DeepSeek V4 ProDeepSeek-V4-Pro1448Independent2026-11-19source
Gemini 3.1 Pro (preview)Gemini 3.1 Pro Preview1446Independent2026-11-19source
Gemini 3.5 Flash-LiteGemini 3.5 Flash-Lite1432Independent2026-11-19source
InklingInkling1414Independent2026-11-19source
Inkling-SmallInkling-Small1408Independent2026-11-19source
Nemotron 3.5 Super VLNVIDIA-Nemotron-3.5-Super-VL-120B-A12B (NVFP4)1402Independentprecision NVFP42026-11-19source
Qwen3.5-397B-A17BQwen3.5-397B-A17B1400Independent2026-11-19source
MiniMax-M2.7MiniMax-M2.71398Independent2026-11-19source
MiniMax-M2.5MiniMax-M2.51386Independent2026-11-19source
Gemma 4 31BGemma 4 31B1370Independent2026-11-19source
DeepSeek-V3.2DeepSeek-V3.2 (Thinking)1362Independentthinking on2026-11-19source
Qwen3.5-122B-A10BQwen3.5-122B-A10B1360Independent2026-11-19source
Gemma 4 26B A4B ITGemma 4 26B A4B1358Independent2026-11-19source
Grok 4.3Grok 4.31357Independent2026-11-19source
Laguna M.1Laguna M.11350Independent2026-11-19source
DeepSeek-V3.2DeepSeek-V3.21326Independent2026-11-19source
Mistral Medium 3.5Mistral Medium 3.51263Independent2026-11-19source
Mistral Large 3Mistral Large 31233Independent2026-11-19source

A rank is among the models we list with a result, not a place on the whole leaderboard, and a model offered by several hosts is shown once, on its own listing. These results don't move our grades. Data: LMArena leaderboard dataset (lmarena-ai/leaderboard-dataset), CC BY 4.0, CC-BY-4.0.

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.