Code Arena | HTML

View overall rankings across AI models on agentic coding tasks involving multi-step reasoning, tool use, and production-style workflows.

Feb 24, 2026

123,082 votes

46 models

	Rank Spread
1	14	claude-opus-4-6 Anthropic · Proprietary	1557+33/-33	419
2	14	claude-opus-4-6-thinking Anthropic · Proprietary	1554+34/-34	367
3	16	gemini-3.1-pro-preview Google · Proprietary	1522+43/-43	222
4	18	claude-sonnet-4-6 Anthropic · Proprietary	1510+35/-35	320
5	36	claude-opus-4-5-20251101-thinking-32k Anthropic · Proprietary	1502+10/-10	8,081
6	310	gpt-5.2-high OpenAI · Proprietary	1477+16/-16	1,696
7	510	claude-opus-4-5-20251101 Anthropic · Proprietary	1469+9/-9	8,265
8	611	gemini-3-pro Google · Proprietary	1461+9/-9	14,273
9	814	glm-4.7 Z.ai · MIT	1447+10/-10	4,991
10	519	glm-5 Z.ai · MIT	1443+33/-33	348
11	914	gemini-3-flash Google · Proprietary	1441+9/-9	9,551
12	618	minimax-m2.5 MiniMax · Modified MIT	1440+28/-28	508
13	919	kimi-k2.5-thinking Moonshot · Modified MIT	1427+22/-22	749
14	923	kimi-k2.5-instant Moonshot · Modified MIT	1418+27/-27	497
15	1119	minimax-m2.1-preview MiniMax · MIT	1413+10/-10	7,292
16	1122	gemini-3-flash (thinking-minimal) Google · Proprietary	1406+10/-10	5,710
17	1123	gpt-5.2 OpenAI · Proprietary	1401+16/-16	1,634
18	1223	gpt-5-medium OpenAI · Proprietary	1399+13/-13	3,929
19	1523	claude-opus-4-1-20250805 Anthropic · Proprietary	1394+9/-9	8,965
20	1523	claude-sonnet-4-5-20250929-thinking-32k Anthropic · Proprietary	1393+8/-8	11,507
21	1524	gpt-5.1-medium OpenAI · Proprietary	1393+9/-9	6,413
22	1624	claude-sonnet-4-5-20250929 Anthropic · Proprietary	1386+8/-8	13,104
23	1128	qwen3.5-397b-a17b Alibaba · Apache 2.0	1386+33/-33	321
24	2128	deepseek-v3.2-thinking DeepSeek · MIT	1373+11/-11	3,858
25	2328	glm-4.6 Z.ai · MIT	1362+9/-9	8,715
26	2329	mimo-v2-flash (non-thinking) Xiaomi · MIT	1361+11/-11	4,328
27	2329	gpt-5.1 OpenAI · Proprietary	1359+8/-8	10,436
28	2331	mimo-v2-flash (thinking) Xiaomi · MIT	1353+18/-18	1,218
29	2631	gpt-5.2-codex OpenAI · Proprietary	1340+12/-12	2,990
30	2831	kimi-k2-thinking-turbo Moonshot · Modified MIT	1339+8/-8	10,056
31	2832	gpt-5.1-codex OpenAI · Proprietary	1334+10/-10	6,489
32	3134	minimax-m2 MiniMax · Apache 2.0	1318+9/-9	8,802
33	3236	claude-haiku-4-5-20251001 Anthropic · Proprietary	1304+8/-8	11,237
34	3236	deepseek-v3.2 DeepSeek · MIT	1302+10/-10	5,018
35	3336	deepseek-v3.2-exp DeepSeek · MIT	1292+11/-11	5,131
36	3336	qwen3-coder-480b-a35b-instruct Alibaba · Apache 2.0	1290+8/-8	10,981
37	3739	KAT-Coder-Pro-V1 KwaiKAT · Proprietary	1265+15/-15	1,955
38	3740	gpt-5.1-codex-mini OpenAI · Proprietary	1249+17/-17	1,537
39	3740	grok-4-1-fast-reasoning xAI · Proprietary	1241+11/-11	5,662
40	3843	mistral-large-3 Mistral · Apache 2.0	1227+20/-20	1,039
41	4043	gemini-2.5-pro Google · Proprietary	1211+13/-13	3,455
42	4043	devstral-2 Mistral · Modified MIT	1211+17/-17	1,453
43	4043	grok-4.1-thinking xAI · Proprietary	1210+19/-19	1,267
44	4445	grok-4-fast-reasoning xAI · Proprietary	1158+23/-23	968
45	4446	grok-code-fast-1 xAI · Proprietary	1146+22/-22	1,017
46	4546	devstral-medium-2507 Mistral · Proprietary	1104+22/-22	1,020

Code Arena | HTML

Remove Style Control Leaderboard Plots

Confidence Intervals on Model Strength (via Bootstrapping)

Battle Count for Each Combination of Models (without Ties)

Fraction of Model A Wins for All Non-tied A vs. B Battles

Average Win Rate Against All Other Models (Uniform Sampling and No Ties)