LiveCodeBench

Artificial Analysis · 50 models

Rank	Model	Score
1	Gemini 3 Pro Preview (High)	0.917
2	Gemini 3 Flash Preview (Reasoning)	0.908
3	DeepSeek V3.2 Speciale	0.896
4	GLM-4.7 (Reasoning)	0.894
4	GPT-5.2 (Medium)	0.894
6	GPT-5.2 (Xhigh)	0.889
7	gpt-oss-120b (High)	0.878
8	Claude Opus 4.5 (Reasoning)	0.871
9	GPT-5.1 (High)	0.868
9	MiMo-V2-Flash (Reasoning)	0.868
11	DeepSeek V3.2 (Reasoning)	0.862
12	o4-mini (High)	0.859
13	Gemini 3 Pro Preview (Low)	0.857
14	Kimi K2 Thinking	0.853
15	GPT-5.1 Codex (High)	0.849
16	GPT-5 (High)	0.846
17	GPT-5 Codex (High)	0.84
18	GPT-5 mini (High)	0.838
19	GPT-5.1 Codex mini (High)	0.836
20	Grok 4 Fast (Reasoning)	0.832
21	MiniMax-M2	0.826
22	Grok 4.1 Fast (Reasoning)	0.822
23	Grok 4	0.819
24	ERNIE 5.0 Thinking Preview	0.812
25	MiniMax-M2.1	0.81
26	o3	0.808
27	Apriel-v1.6-15B-Thinker	0.807
28	Gemini 2.5 Pro	0.801
29	DeepSeek V3.1 Terminus (Reasoning)	0.798
30	Gemini 3 Flash Preview (Non-reasoning)	0.797
31	DeepSeek V3.2 Exp (Reasoning)	0.789
31	GPT-5 nano (High)	0.789
33	Qwen3 235B A22B 2507 (Reasoning)	0.788
34	DeepSeek V3.1 (Reasoning)	0.784
34	Qwen3 Next 80B A3B (Reasoning)	0.784
36	Gemini 2.5 Pro Preview (Mar' 25)	0.778
37	INTELLECT-3 (Based on GLM-4.5-Air)	0.777
37	gpt-oss-20b (High)	0.777
39	DeepSeek R1 0528 (May '25)	0.77
39	Gemini 2.5 Pro Preview (May' 25)	0.77
41	K-EXAONE (Reasoning)	0.768
42	Qwen3 Max	0.767
43	Doubao Seed Code	0.766
44	Seed-OSS-36B-Instruct	0.765
45	GPT-5 (Low)	0.763
45	GPT-5 nano (Medium)	0.763
47	Magistral Medium 1.2	0.75
48	EXAONE 4.0 32B (Reasoning)	0.747
48	KAT-Coder-Pro V1	0.747
50	NVIDIA Nemotron 3 Nano 30B A3B (Reasoning)	0.741