Benchmarks / Coding

LiveCodeBench (Vals)

Our Implementation of the LiveCodeBench benchmark Scores are the values published by VALS AI. Source benchmark version 1.

125models
–Open Index weight
2026-09-01last updated
Vals AIpublisher

LiveCodeBench (Vals)

Higher is better · best effort per model

Chart settings
125 of 125 selected

full table below

#ModelEffortScoreCost / taskHarness
1Fable 5.1max90.5%$0.261–
2Fable 5max89.8%$0.428–
3Gemini 3.8 Flashhigh89.5%$0.071–
4Opus 5max89.0%$0.134–
5Gemini 3.7 Flashhigh88.7%$0.040–
6Gemini 3.1 Prohigh88.5%$0.101–
7Grok 4.6high88.2%$0.040–
8Gemini 3.6 Flashhigh88.1%$0.056–
9GPT-5.2 Codexxhigh88.0%$0.117–
10Qwen 3.8 Maxmax87.9%$0.091–
11Opus 4.8max87.8%$0.327–
12Gemini 3.5 Flashhigh87.6%$0.086–
13DeepSeek V4 Pro 0813max87.5%$0.064–
14DeepSeek V4max87.5%$0.110–
15Grok 4.5high87.4%$0.045–
16GPT-5.3 Codexxhigh87.3%$0.092–
17DeepSeek V4 Flash 0731high87.3%$0.020–
18Kimi K3max87.2%$0.104–
19Qwen 3.7 Maxmax87.1%$0.087–
20Kimi K2.6max86.8%$0.065–
21GPT-5 Minihigh86.6%$0.012–
22GPT-5.1high86.5%$0.126–
23Gemini 3 Prohigh86.4%$0.127–
24Nemotron 3 Ultra 550B A55Bmax86.0%––
25Qwen 3.6 Plusmax86.0%$0.078–
26Inkling Smallxhigh85.9%$0.021–
27GPT-5.6 Terraxhigh85.9%$0.031–
28GPT-5high85.9%$0.064–
29Muse Spark 1.1xhigh85.9%$0.046–
30Gemini 3 Flashhigh85.6%$0.051–
31GPT-5.1 Codexmax85.5%$0.138–
32Inklingmax85.5%$0.153–
33GPT-5.2xhigh85.4%$0.135–
34Qwen 3.5 Plusnone85.3%$0.048–
35GPT-5.5xhigh85.3%$0.067–
36Opus 4.7max85.1%$0.351–
37GPT-5 Codexhigh84.7%$0.097–
38Opus 4.6max84.7%$0.343–
39Grok 4.3high84.5%$0.020–
40Grok 4.20xhigh84.3%$0.052–
41GPT-5.4xhigh84.1%$0.329–
42GPT-5.4 Nanohigh84.0%$0.0025–
43Qwen 3.8 27Bxhigh84.0%$0.045–
44Ling 3.0 Flash–84.0%$0.0030–
45o3high83.9%$0.033–
46Kimi K2.5max83.9%$0.039–
47Opus 4.5high83.7%$0.971–
48Qwen 3.5 Flashmax83.3%$0.0095–
49Grok 4high83.2%$0.156–
50gpt-oss-120bhigh83.2%$0.0085–
51GPT-5.6 Solmax82.6%$0.090–
52Sonnet 5max82.4%$0.095–
53GLM 4.7none82.2%$0.035–
54o4 minihigh82.2%$0.014–
55MiniMax M3max82.2%$0.037–
56Sonnet 4.6max82.1%$0.371–
57Kimi K2.7 Codemax82.0%$0.027–
58GLM 5high81.9%$0.044–
59MiniMax M2.1–81.8%$0.018–
60MiMo V2.5xhigh81.5%$0.0022–
61GPT-5.4 Minixhigh81.5%$0.033–
62GLM 5.1max81.4%$0.046–
63MiMo V2.5 Proxhigh81.4%$0.0068–
64GLM 4.6none81.0%$0.027–
65DeepSeek V3.2high80.7%$0.0033–
66Grok 4.1 Fasthigh80.6%$0.0039–
67GLM 5.3max80.5%$0.069–
68GLM 5.3 Flashmax80.5%$0.0071–
69gpt-oss-20bhigh80.4%$0.0067–
70Gemini 3.1 Flash-Litehigh80.1%$0.0043–
71MiniMax M2.7high79.9%$0.021–
72MiniMax M2.5high79.2%$0.033–
73Gemini 2.5 Prohigh79.2%$0.020–
74Gemini 3.5 Flash-Litehigh79.0%$0.019–
75Grok 4 Fastnone79.0%$0.0045–
76Qwen 3 Maxmax78.2%$0.021–
77Grok 3 Minihigh76.2%$0.0081–
78Gemini 2.5 Flashmedium76.2%$0.0024–
79Magistral Medium 1.2medium74.9%$0.050–
80Sonnet 4.5high73.0%$0.135–
81Magistral Small 1.2–72.1%$0.015–
82o3 minihigh71.5%$0.033–
83Gemini 2.5 Flash Litenone71.4%$0.0004–
84Qwen 3 235B–70.6%$0.011–
85Kimi K2medium70.4%$0.0069–
86DeepSeek R1xhigh70.2%$0.041–
87GPT-5 Nanohigh70.2%$0.0051–
88Opus 4medium70.2%$0.397–
89GLM 5.2max69.5%$0.129–
90Laguna M.1–68.1%$0.0057–
91Laguna XS.2–67.8%$0.0023–
92GLM 4.5xhigh67.4%$0.018–
93Opus 4.1high66.5%$0.361–
94DeepSeek V3none65.5%$0.0018–
95Sonnet 4high62.4%$0.042–
96Grok Code Fast–62.0%$0.0085–
97Sonnet 3.7medium60.4%$0.125–
98Llama 3.3none58.4%––
99GPT-4.1 Minihigh58.2%$0.0012–
100Mistral Large 3none55.3%$0.0022–
101GPT-4.1high54.7%$0.0093–
102Grok 3none52.9%$0.0094–
103Devstral 2none51.8%$0.057–
104o1high50.3%$0.328–
105Claude 3.5 Sonnet Latest–49.6%$0.012–
106Llama 4 Maverickmedium47.3%$0.0008–
107Mistral Medium 3.1none44.8%$0.0016–
108Gemini 2.0 Flashmedium43.6%$0.0003–
109GPT-4ohigh43.4%$0.0040–
110Devstral Small 2none43.2%$0.0003–
111GPT-4.1 Nanohigh42.7%$0.0003–
112Haiku 3.5medium41.9%$0.0031–
113Gemini 1.5 Pronone41.7%$0.0020–
114Haiku 4.5xhigh41.2%$0.060–
115Grok 2none38.7%$0.0030–
116Llama 4 Scoutmedium38.5%$0.0006–
117Mistral Large 11–37.1%$0.0033–
118Gemini 1.5 Flashnone36.9%$0.0001–
119Command Anone35.1%$0.0066–
120Mistral Small 3.1none31.8%$0.0001–
121GPT-4o minihigh26.4%$0.0003–
122Jamba 1.6 Largenone22.3%$0.0032–
123Command-R+ (Apr '24)none18.2%$0.0033–
124Mistral Small 02–15.8%$0.0004–
125Jamba 1.6 Mininone9.9%$0.0002–