LiveCodeBench

Competitive programming problems released after model cutoffs. Run by Vals AI.

Published by

Vals AI

Category

Coding

Index weight

1.0

Models

135

Data as of

9 Sept 2026

Vals AI (vals.ai).

1. 1Claude Fable 5.190.5%
2. 2Claude Fable 589.8%
3. 3Gemini 3.8 Flash89.5%
4. 4Claude Opus 589.0%
5. 5Gemini 3.7 Flash88.7%
6. 6Gemini 3.1 Pro88.5%
7. 7Grok 4.688.2%
8. 8Gemini 3.6 Flash88.1%
9. 9GPT-5.2 Codex88.0%
10. 10Qwen3 887.8%
11. 11Claude Opus 4.887.8%
12. 12Gemini 3.5 Flash87.6%
13. 13DeepSeek V4 Pro87.5%
14. 14Grok 4.587.3%
15. 15GPT-5.3 Codex87.3%

SearchAll providersAI21 LabsAlibabaAnthropicCohereDeepSeekGoogleInclusionAIMetaMiniMaxMistral AIMoonshot AINVIDIAOpenAIPoolsideThinking MachinesUnknownxAIXiaomiZhipu AIOpen weights135 of 135

#	Model	Score ↓	Run ↕	Index ↕	Published ↕
1	Claude Fable 5.1Anthropic	90.5%	–	70.0	–
2	Claude Fable 5Anthropic	89.8%	–	70.4	–
3	Gemini 3.8 FlashGoogle	89.5%	–	60.1	–
4	Claude Opus 5Anthropic	89.0%	–	70.0	–
5	Gemini 3.7 FlashGoogle	88.7%	–	60.8	–
6	Gemini 3.1 ProGoogle	88.5%	–	60.7	–
7	Grok 4.6xAI	88.2%	–	58.7	–
8	Gemini 3.6 FlashGoogle	88.1%	–	58.6	–
9	GPT-5.2 CodexOpenAI	88.0%	–	60.7	–
10	Qwen3 8Alibaba	87.8%	–	63.9	–
11	Claude Opus 4.8Anthropic	87.8%	–	64.7	–
12	Gemini 3.5 FlashGoogle	87.6%	–	60.4	–
13	DeepSeek V4 ProDeepSeekopen	87.5%	max, 0813	56.1	–
14	Grok 4.5xAI	87.3%	–	54.2	–
15	GPT-5.3 CodexOpenAI	87.3%	–	–	–
16	DeepSeek V4 FlashDeepSeekopen	87.3%	high, 0731	59.3	–
17	Kimi K3Moonshot AIopen	87.2%	–	62.7	–
18	Qwen3 7Alibaba	87.1%	–	58.9	–
19	Kimi K2.6Moonshot AIopen	86.8%	kimi/kimi-k2.6	60.6	–
20	GPT-5 miniOpenAI	86.6%	–	52.3	–
21	GPT-5.1OpenAI	86.5%	–	53.8	–
22	Gemini 3 ProGoogle	86.4%	–	61.6	–
23	Nemotron 3 Ultra 550B A55BNVIDIAopen	86.0%	–	58.3	–
24	Qwen3.6 PlusAlibabaopen	86.0%	–	59.0	–
25	GPT-5.6 TerraOpenAI	85.9%	–	64.2	–
26	TInkling-SmallThinking Machinesopen	85.9%	–	55.0	–
27	Muse Spark 1.1Meta	85.9%	–	56.4	–
28	GPT-5OpenAI	85.9%	–	54.5	–
29	Gemini 3 FlashGoogle	85.6%	–	58.5	–
30	GPT-5.1-CodexOpenAI	85.5%	–	57.1	–
31	GPT-5.2OpenAI	85.4%	–	57.3	–
32	Qwen3.5 PlusAlibaba	85.3%	–	57.2	–
33	GPT-5.5OpenAI	85.3%	–	63.7	–
34	Claude Opus 4.7Anthropic	85.1%	–	65.5	–
35	GPT-5-CodexOpenAI	84.7%	–	–	–
36	Claude Opus 4.6Anthropic	84.7%	–	62.3	–
37	Grok 4.3xAI	84.5%	–	50.1	–
38	Grok 4.20xAI	84.3%	–	54.3	–
39	GPT-5.4OpenAI	84.1%	–	61.3	–
40	GPT-5.4 nanoOpenAI	84.0%	–	48.3	–
41	Qwen 3.8 27BAlibabaopen	84.0%	–	52.0	–
42	ILing 3.0 FlashInclusionAIopen	84.0%	2607	54.0	–
43	o3OpenAI	83.9%	–	52.9	–
44	Kimi K2.5Moonshot AIopen	83.9%	–	55.3	–
45	Claude Opus 4.5Anthropic	83.7%	–	60.6	–
46	GPT-5.1-Codex-MaxOpenAI	83.6%	–	–	–
47	Qwen3.5 FlashAlibaba	83.3%	–	51.8	–
48	Grok 4xAI	83.3%	0709	57.2	–
49	gpt-oss-120bOpenAIopen	83.2%	–	47.6	–
50	GPT-5.6 SolOpenAI	82.6%	–	65.5	–
51	Claude Sonnet 5Anthropic	82.4%	–	61.2	–
52	GLM-4.7Zhipu AIopen	82.2%	zai/glm-4.7	54.9	–
53	o4-miniOpenAI	82.2%	–	48.0	–
54	MiniMax-M3MiniMaxopen	82.2%	minimax/MiniMax-M3	56.9	–
55	Claude Sonnet 4.6Anthropic	82.1%	–	60.3	–
56	Kimi K2.7 CodeMoonshot AIopen	82.0%	–	56.2	–
57	GLM-5Zhipu AIopen	81.9%	–	57.0	–
58	MiniMax-M2MiniMaxopen	81.8%	minimax/MiniMax-M2.1	55.1	–
59	GPT-5.4 miniOpenAI	81.5%	–	46.0	–
60	GLM-5.1Zhipu AIopen	81.4%	zai/glm-5.1	60.1	–

Show 75 more