ELO Rating78 models ranked

Coding ELO Leaderboard 2026

Coding ELO is a specialised Chatbot Arena leaderboard restricted to coding tasks. It measures human preference when comparing model-generated code, debugging explanations, and programming help.

Quick Answer

The best model on Coding ELO in 2026 is Claude Opus 4 by Anthropic, scoring 1497 ELO. Runner-up: Claude Sonnet 4 (1473).

Provider

78 / 78 models

#	Model	Provider	Score	Percentile
🥇	Claude Opus 4	Anthropic	1497ELO	99th
🥈	Claude Sonnet 4	Anthropic	1473ELO	97th
🥉	Gemini 2.5 Pro	Google	1445ELO	96th
4	DeepSeek R1	DeepSeek	1398ELO	95th
5	Qwen 2.5 Max	Alibaba	1374ELO	94th
6	Gemini 2.0 Flash	Google	1360ELO	92th
7	DeepSeek V3	DeepSeek	1358ELO	91th
8	o3-mini	OpenAI	1348ELO	90th
9	o3	OpenAI	1320ELO	88th
10	o1	OpenAI	1310ELO	87th
11	Claude 3.5 Sonnet	Anthropic	1300ELO	86th
12	GPT-4.5	OpenAI	1295ELO	85th
13	Qwen 3 235B MoE	Alibaba	1290ELO	83th
14	Qwen 2.5 Coder 32B	Alibaba	1290ELO	82th
15	Gemini Experimental 1206	Google	1280ELO	81th
16	Llama 4 Maverick	Meta	1280ELO	79th
17	Grok 3	xAI	1280ELO	78th
18	o1-mini	OpenAI	1280ELO	77th
19	DeepSeek R1 (Groq)	Groq	1270ELO	76th
20	DeepSeek R1 (Together)	Together AI	1270ELO	74th
21	Gemini 2.0 Flash Thinking	Google	1270ELO	73th
22	ChatGPT-4o Latest	OpenAI	1270ELO	72th
23	o4-mini	OpenAI	1270ELO	71th
24	GPT-4o (Aug 2024)	OpenAI	1265ELO	69th
25	Gemini 2.5 Flash	Google	1260ELO	68th
26	QwQ 32B	Alibaba	1250ELO	67th
27	Claude 3.5 Haiku	Anthropic	1250ELO	65th
28	Codestral 22B	Mistral AI	1250ELO	64th
29	GPT-4 Turbo	OpenAI	1245ELO	63th
30	DeepSeek R1 Distill Llama 70B	DeepSeek	1240ELO	62th
31	Mistral Large	Mistral	1240ELO	60th
32	DeepSeek R1 Distill Qwen 32B	DeepSeek	1235ELO	59th
33	Sonar Reasoning	Perplexity	1235ELO	58th
34	Llama 4 Scout	Meta	1230ELO	56th
35	Command A	Cohere	1230ELO	55th
36	Grok 2	xAI	1225ELO	54th
37	DeepSeek V2.5	DeepSeek	1220ELO	53th
38	Mixtral 8x22B (Fireworks)	Fireworks AI	1220ELO	51th
39	WizardLM-2 8x22B	Microsoft	1220ELO	50th
40	Qwen 2.5 72B	Alibaba	1210ELO	49th
41	Qwen 2.5 72B (Together)	Together AI	1210ELO	47th
42	Amazon Nova Pro	Amazon	1210ELO	46th
43	GPT-4 1	OpenAI	1210ELO	45th
44	Llama 3.1 405B (Fireworks)	Fireworks AI	1200ELO	44th
45	Llama 3.1 405B	Meta	1200ELO	42th
46	Llama 3.1 405B (Together)	Together AI	1200ELO	41th
47	GPT-4 1.5-mini	OpenAI	1200ELO	40th
48	Claude Haiku 4	Anthropic	1195ELO	38th
49	Sonar Pro	Perplexity	1195ELO	37th
50	Gemma 2 27B	Google	1195ELO	36th
51	Yi-Large	01.AI	1195ELO	35th
52	Phi-3.5 MoE	Microsoft	1190ELO	33th
53	Grok 3-mini	xAI	1190ELO	32th
54	Llama 3.3 70B (Fireworks)	Fireworks AI	1180ELO	31th
55	Llama 3.3 70B (Groq)	Groq	1180ELO	29th
56	Llama 3.3 70B	Meta	1180ELO	28th
57	Llama 3.3 70B (Together)	Together AI	1180ELO	27th
58	Phi-3 Medium	Microsoft	1175ELO	26th
59	Gemini 2.0 Flash Lite	Google	1170ELO	24th
60	Yi-Lightning	01.AI	1170ELO	23th
61	Sonar	Perplexity	1170ELO	22th
62	Phi-3.5 Mini	Microsoft	1165ELO	21th
63	GPT-4o	OpenAI	1162ELO	19th
64	Command R+	Cohere	1160ELO	18th
65	Mistral Small	Mistral	1160ELO	17th
66	Amazon Nova Lite	Amazon	1160ELO	15th
67	Gemma 2 9B (Groq)	Groq	1160ELO	14th
68	GPT-4 1.5-nano	OpenAI	1160ELO	13th
69	Gemma 2 9B	Google	1155ELO	12th
70	Mixtral 8x7B (Groq)	Groq	1150ELO	10th
71	InternLM 2.5 20B	Shanghai AI Lab	1150ELO	9th
72	Phi-4	Microsoft	1130ELO	8th
73	Command R	Cohere	1100ELO	6th
74	Amazon Nova Micro	Amazon	1100ELO	5th
75	Command R7B	Cohere	1100ELO	4th
76	GPT-3.5 Turbo	OpenAI	1100ELO	3th
77	GPT-4o Mini	OpenAI	1098ELO	1th
78	Mistral 7B (Together)	Together AI	1090ELO	0th

What Coding ELO Tests

Human preference on coding-specific tasks: writing functions, explaining bugs, reviewing pull requests. Derived from the same Arena methodology applied only to code conversations.

Score Range

1100–1450+ (average ~1220)

Source

LMSYS Chatbot Arena — Coding ↗

Other Benchmarks

Arena ELO Reasoning ELO HumanEval MMLU MATH GPQA

Compare models side-by-side

Full spec comparison — pricing, context window, and all benchmarks.

Compare Models →