benchgap
Coding

Vals LiveCodeBench leaderboard

As of 2026-11-21, the highest measured score on Vals LiveCodeBench is 90.5% by Claude Fable 5.1. 165 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#ModelScoreSource
1Gemini 4 Carbon94.8%estimated ± 3.0 pp, medium confidence
2Claude Fable 5.294.6%estimated ± 3.0 pp, medium confidence
3Gemini 4 Hydrogen91.5%estimated ± 3.0 pp, medium confidence
4Gemini 4 Argon90.8%estimated ± 3.0 pp, medium confidence
5Ember-190.8%estimated ± 10.0 pp, low confidence
6Claude Fable 5.190.5%measured
7Claude Fable 589.8%measured
8Muse Wonder Full89.8%estimated ± 5.7 pp, low confidence
9Gemini 3.8 Flash89.5%measured
10Muse Wonder89.4%estimated ± 3.0 pp, high confidence
11Claude Opus 5.589.3%estimated ± 1.9 pp, medium confidence
12MiMo-V2.6-Pro89.0%estimated ± 5.7 pp, low confidence
13Claude Opus 589.0%measured
14Kimi K3.1 Pro89.0%estimated ± 5.7 pp, low confidence
15GPT-6 Astra89.0%estimated ± 1.9 pp, high confidence
16Sakana Fugu88.8%estimated ± 0.8 pp, low confidence
17Gemini 3.7 Flash88.7%measured
18Gemini 3.1 Pro88.5%measured
19Step 5 Preview88.5%estimated ± 0.8 pp, low confidence
20Sakana Fugu-Ultra88.4%estimated ± 0.8 pp, low confidence
21Claude Mythos 588.3%estimated ± 4.7 pp, medium confidence
22Grok 4.688.2%measured
23Gemini 3.6 Flash88.1%measured
24SWE-288.1%estimated ± 1.9 pp, high confidence
25GPT-5.2-Codex88.0%measured
26DeepSeek V4.1 Flash 101788.0%estimated ± 3.0 pp, high confidence
27Composer 2.587.9%estimated ± 2.8 pp, low confidence
28Qwen3.8 Max87.9%measured
29Muse Spark Open87.9%estimated ± 5.7 pp, low confidence
30Claude Opus 4.887.8%measured
31Gemini 3.5 Flash87.6%measured
32DeepSeek V4 Pro87.5%measured
33GPT-6 Sol87.5%estimated ± 5.7 pp, low confidence
34Grok 4.587.4%measured
35Muse Spark 1.287.4%estimated ± 5.7 pp, low confidence
36DeepSeek V4 Flash 073187.3%measured
37GPT-5.3 Codex87.3%measured
38Grok 4.787.2%estimated ± 3.0 pp, high confidence
39Kimi K387.2%measured
40Qwen3.7 Max87.1%measured
41Claude Haiku 5.587.1%estimated ± 1.9 pp, high confidence
42Ling 3.1 Flash87.1%estimated ± 4.7 pp, medium confidence
43Claude Sonnet 5.587.0%estimated ± 1.9 pp, high confidence
44SWE-2 Flash86.9%estimated ± 1.9 pp, high confidence
45Kimi K2.686.8%measured
46Qwen3.7 Plus86.6%estimated ± 0.8 pp, medium confidence
47GPT-5.186.5%measured
48Hy4 preview86.5%estimated ± 4.7 pp, medium confidence
49Gemini 3 Pro86.4%measured
50Ornith-1.5-397B86.3%estimated ± 4.7 pp, medium confidence
51GPT-6.1 Terra86.2%estimated ± 5.7 pp, low confidence
52Beam 1.0 501B-A23B86.1%estimated ± 0.8 pp, medium confidence
53Qwen3.8-Omni-Flash86.1%estimated ± 3.4 pp, low confidence
54Solar Open 286.1%estimated ± 3.4 pp, low confidence
55Qwen3.8-Flash-Next86.1%estimated ± 3.4 pp, low confidence
56dots3-note Preview86.1%estimated ± 3.4 pp, medium confidence
57Ternary Bonsai 2 27B86.1%estimated ± 3.4 pp, medium confidence
58Nemotron 3.5 Super VL 120B A12B86.1%estimated ± 3.4 pp, medium confidence
59BTL-386.1%estimated ± 3.4 pp, medium confidence
60MAI-Thinking-186.1%estimated ± 3.4 pp, medium confidence
61Nemotron 3.5 Super VL 120B A12B NVFP486.0%estimated ± 3.4 pp, medium confidence
62Nemotron 3 Ultra86.0%measured
63Qwen3.6 Plus86.0%measured
64Kimi K2.586.0%estimated ± 0.8 pp, medium confidence
65SWE-1.786.0%estimated ± 1.9 pp, high confidence
66GPT-6 Luna86.0%estimated ± 5.7 pp, low confidence
67GPT-585.9%measured
68GPT-5.6 Terra85.9%measured
69Inkling-Small85.9%measured
70Muse Spark 1.185.9%measured
71Beam 1.0 501B-A23B FP885.9%estimated ± 0.8 pp, medium confidence
72GPT-6.1 Sol85.8%estimated ± 5.7 pp, low confidence
73Step 3.5 Flash85.7%estimated ± 3.4 pp, medium confidence
74Gemini 3 Flash85.6%measured
75GPT-5.1-Codex85.6%measured
76Qwen 3.6 Max (preview)85.6%estimated ± 0.8 pp, medium confidence
77Beam 1.0 501B-A23B NVFP485.6%estimated ± 0.8 pp, medium confidence
78Inkling85.5%measured
79Ornith-1.0-397B85.5%estimated ± 4.7 pp, medium confidence
80GPT-5.285.4%measured
81MiniMax M3.1 Flash85.3%estimated ± 5.7 pp, low confidence
82GPT-5.585.3%measured
83Claude Opus 4.785.1%measured
84Sage-A1.584.9%estimated ± 0.8 pp, medium confidence
85Muse Glimmer 30B84.8%estimated ± 0.8 pp, medium confidence
86Inkling 2 Small84.7%estimated ± 5.7 pp, low confidence
87Muse Spark 1.384.7%estimated ± 3.0 pp, high confidence
88Claude Opus 4.6 (Adaptive)84.7%measured
89Qwen3.8 Max Preview84.6%estimated ± 5.7 pp, low confidence
90Grok 4.384.5%measured
91Atria Dawn Preview84.5%estimated ± 4.7 pp, medium confidence
92Ornith-1.5-35B-A3B84.5%estimated ± 4.7 pp, medium confidence
93Laguna S 2.184.4%estimated ± 4.7 pp, medium confidence
94Grok 4.2084.3%measured
95Hy3 Preview84.2%estimated ± 0.8 pp, low confidence
96MiMo-V2.6-Flash84.2%estimated ± 5.7 pp, low confidence
97A.X K284.1%estimated ± 0.8 pp, low confidence
98GPT-5.484.1%measured
99GPT-5.4 nano84.0%measured
100Ling 3.0 Flash84.0%measured
101Qwen3.8-27B84.0%measured
102Ling 3.0 Flash FP884.0%estimated ± 0.8 pp, low confidence
103o383.9%measured
104Ripple-65B Preview83.7%estimated ± 0.8 pp, low confidence
105Claude Opus 4.5 Thinking83.7%measured
106GPT-5.1-Codex-Max83.6%measured
107Granite 4.2 30B83.6%estimated ± 0.8 pp, low confidence
108Mercury 2.583.4%estimated ± 0.8 pp, low confidence
109Qwen3.5 Flash83.3%measured
110K-EXAONE 2.083.2%estimated ± 0.8 pp, low confidence
111Grok 483.2%measured
112Spark-X2.6-4B83.1%estimated ± 0.8 pp, low confidence
113Granite 4.2 8B82.9%estimated ± 0.8 pp, low confidence
114Step 3.7 Flash82.7%estimated ± 4.7 pp, medium confidence
115GPT-5.6 Sol82.6%measured
116Hy382.6%estimated ± 5.7 pp, low confidence
117Claude Sonnet 582.4%measured
118Quasar 438B82.3%estimated ± 5.7 pp, low confidence
119Solar Pro 482.2%estimated ± 5.7 pp, low confidence
120GLM-4.782.2%measured
121MiniMax M382.2%measured
122o4-mini82.2%measured
123Claude Sonnet 4.682.1%measured
124Kimi K2.7 Code82.1%measured
125LFM2.5-24B-A2B82.1%estimated ± 0.8 pp, low confidence
126GLM-581.9%estimated ± 4.7 pp, medium confidence
127Nemotron 3 Nano Omni 30B A3B81.9%estimated ± 0.8 pp, low confidence
128Claude Opus 4.681.8%estimated ± 1.9 pp, high confidence
129Nemotron 3.5 Lightning 30B A3B NVFP481.8%estimated ± 0.8 pp, low confidence
130GPT-5.6 Luna81.7%estimated ± 3.0 pp, high confidence
131GPT-5.4 mini81.5%measured
132MiMo-V2.581.5%measured
133GLM-5.181.4%measured
134MiMo-V2.5-Pro81.4%measured
135Agents-A1-4B81.3%estimated ± 0.8 pp, low confidence
136Gemini 2.5 Pro81.1%estimated ± 5.7 pp, low confidence
137GLM-4.681.0%measured
138Inkling 2 Mini80.7%estimated ± 10.0 pp, low confidence
139Grok 4.1 Fast (Reasoning)80.6%measured
140Apodex 1.180.6%estimated ± 5.7 pp, low confidence
141Apodex 1.1 Mini80.6%estimated ± 5.7 pp, low confidence
142GLM-5.380.5%measured
143GLM-5.3-Flash80.5%measured
144Gemini 3.1 Flash-Lite80.1%measured
145BTL-479.9%estimated ± 10.0 pp, low confidence
146MiniMax M2.779.9%measured
147Ling 3.0 Flash VL79.7%estimated ± 5.7 pp, low confidence
148Muse Spark79.7%estimated ± 4.7 pp, medium confidence
149MiMo-V2-Pro79.5%estimated ± 10.0 pp, low confidence
150Gemini 3.5 Flash-Lite79.0%measured
151Grok 4 Fast (Reasoning)79.0%measured
152Claude Sonnet 4.578.7%estimated ± 10.0 pp, low confidence
153Qwen3 Max78.2%measured
154Ornith-1.0-35B77.5%estimated ± 4.7 pp, medium confidence
155ERNIE 5.5 Flash Preview77.4%estimated ± 6.1 pp, low confidence
156Hy477.4%estimated ± 6.1 pp, low confidence
157Hy4 flash77.4%estimated ± 6.1 pp, low confidence
158LongCat-2.577.4%estimated ± 6.1 pp, low confidence
159Qwen4 Max77.4%estimated ± 6.1 pp, low confidence
160Trinity-Large-Preview77.1%estimated ± 5.7 pp, low confidence
161Trinity-Large-Thinking77.1%estimated ± 5.7 pp, low confidence
162Mistral Medium 3.5 128B76.8%estimated ± 5.7 pp, low confidence
163Gemma 4 26B A4B76.7%estimated ± 5.7 pp, low confidence
164Nemotron 3.5 Bolt 74B-A9B76.4%estimated ± 5.7 pp, low confidence
165MiMo-V2-Omni76.1%estimated ± 10.0 pp, low confidence
166DeepSeek V3 032475.9%estimated ± 5.7 pp, low confidence
167Claude 4.1 Opus75.8%estimated ± 10.0 pp, low confidence
168GPT-OSS 20B75.8%estimated ± 5.7 pp, low confidence
169Mistral Small 475.7%estimated ± 5.7 pp, low confidence
170North Mini Code75.7%estimated ± 5.7 pp, low confidence
171Command A+75.5%estimated ± 5.7 pp, low confidence
172Claude Opus 4.575.0%measured
173Jais-3-MoVA-36B-A4B73.9%estimated ± 10.0 pp, low confidence
174Claude 4 Sonnet73.8%estimated ± 10.0 pp, low confidence
175MAI-Code-1.1-Flash73.7%estimated ± 10.0 pp, low confidence
176Laguna XS 2.173.6%estimated ± 4.7 pp, medium confidence
177Claude 4 Opus73.6%estimated ± 10.0 pp, low confidence
178Ornith-1.5-9B73.5%estimated ± 4.7 pp, medium confidence
179Jais-3-32B73.5%estimated ± 10.0 pp, low confidence
180DeepSeek V373.2%estimated ± 5.7 pp, low confidence
181Claude Sonnet 4.5 Thinking73.0%measured
182GPT-OSS 120B71.5%estimated ± 5.7 pp, low confidence
183o3-mini71.5%measured
184Jais-3-375B-A23B71.0%estimated ± 10.0 pp, low confidence
185Jais-3-7B71.0%estimated ± 10.0 pp, low confidence
186Qwen3-Coder 480B-A35B70.3%estimated ± 10.0 pp, low confidence
187Qwen3.5-35B-A3B69.8%estimated ± 10.0 pp, low confidence
188GLM-5.269.5%measured
189Llama 4 Maverick69.3%estimated ± 5.7 pp, low confidence
190o1-preview68.8%estimated ± 7.1 pp, medium confidence
191Laguna M.168.1%measured
192Laguna XS.267.8%measured
193GLM-4.567.4%measured
194K-Exaone67.1%estimated ± 7.1 pp, medium confidence
195DeepSeek V3.166.1%estimated ± 10.0 pp, low confidence
196Gemma 4 12B66.1%estimated ± 7.1 pp, medium confidence
197Kimi K265.9%estimated ± 10.0 pp, low confidence
198Mistral Small 3.165.1%estimated ± 5.7 pp, low confidence
199Ornith-1.0-9B64.0%estimated ± 4.7 pp, low confidence
200Solar Pro 362.3%estimated ± 5.7 pp, low confidence
201Granite 4.2 3B62.1%estimated ± 5.7 pp, low confidence
202Grok Code Fast 162.0%measured
203Gemma 4 E4B60.9%estimated ± 5.7 pp, low confidence
204Ling 3.0 Tiny60.6%estimated ± 5.7 pp, low confidence
205Claude 3.7 Sonnet Thinking60.4%measured
206Ling 2.6 Flash60.4%estimated ± 7.1 pp, medium confidence
207Nemotron 3 Super 100B60.2%estimated ± 10.0 pp, low confidence
208Gemma 3 27B59.4%estimated ± 5.7 pp, low confidence
209GLM-4.7-Flash59.0%estimated ± 10.0 pp, low confidence
210GPT-4.1 mini58.2%measured
211LongCat-Flash-Lite-Sparse57.8%estimated ± 4.7 pp, low confidence
212DeepSeek-R1-052857.6%estimated ± 10.0 pp, low confidence
213Celeris-157.0%estimated ± 5.7 pp, low confidence
214Claude 3.7 Sonnet56.7%measured
215Llama 4 Scout56.6%estimated ± 5.7 pp, low confidence
216MiniMax M1 80k56.2%estimated ± 10.0 pp, low confidence
217GPT-4 Turbo55.8%estimated ± 7.1 pp, medium confidence
218Mistral Large 355.3%measured
219GPT-4.154.7%measured
220ZAYA1-74B-Preview54.1%estimated ± 10.0 pp, low confidence
221Claude 3 Opus53.1%estimated ± 7.1 pp, medium confidence
222Qwen3.6-27B52.0%estimated ± 3.4 pp, low confidence
223DeepSeek-R151.7%estimated ± 10.0 pp, low confidence
224Devstral Small50.6%estimated ± 10.0 pp, low confidence
225o150.3%measured
226Sarvam 105B49.9%estimated ± 10.0 pp, low confidence
227Claude 3.5 Sonnet49.6%measured
228LFM3-1.7B49.4%estimated ± 10.0 pp, low confidence
229Sarvam 30B47.7%estimated ± 10.0 pp, low confidence
230Gemini 2.5 Flash-Lite47.5%estimated ± 10.0 pp, low confidence
231Gemini 2.0 Flash-Lite47.1%estimated ± 10.0 pp, low confidence
232LFM2.5-2.6B44.6%estimated ± 5.7 pp, low confidence
233Nemotron 3 Nano 30B44.5%estimated ± 7.1 pp, medium confidence
234Gemini 2.0 Flash43.6%measured
235GPT-4o43.4%measured
236GPT-4.1 nano42.7%measured
237Claude 3.5 Haiku41.9%measured
238Gemini 1.5 Pro41.7%measured
239Claude Haiku 4.541.2%measured
240Ultravox v0.6 Llama 3.3 70B39.6%estimated ± 7.1 pp, medium confidence
241Qwen3.5 397B39.2%estimated ± 3.4 pp, low confidence
242Gemini 1.5 Flash36.9%measured
243Gemma 4 E2B27.8%estimated ± 7.1 pp, low confidence
244GPT-4o mini26.4%measured
245LLaDA2.2-flash20.8%estimated ± 4.7 pp, low confidence
246Mellum2.1-12B-A2.5B-Thinking2.8%estimated ± 3.4 pp, low confidence
247MiniCPM5-2B0.3%estimated ± 4.7 pp, low confidence
248Qwen3.5-27B0.2%estimated ± 3.4 pp, low confidence
249MiMo-V2-Flash0.2%estimated ± 3.4 pp, low confidence
250Qwen3.6-35B-A3B0.1%estimated ± 3.4 pp, low confidence
251Flint0.1%estimated ± 3.4 pp, low confidence
252Gemma 4 31B0.1%estimated ± 3.4 pp, low confidence
253Qwen3.5-122B-A10B0.0%estimated ± 3.4 pp, low confidence

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General