benchgap
Coding

PostTrainBench v1.1 leaderboard

As of 2026-11-18, the highest measured score on PostTrainBench v1.1 is 49.3% by Claude Opus 5.5. 170 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#ModelScoreSource
1Gemini 4 Carbon49.3%estimated ± 1.7 pp, low confidence
2Claude Fable 5.249.3%estimated ± 1.7 pp, low confidence
3Claude Opus 5.549.3%measured
4Claude Sonnet 5.549.2%estimated ± 1.7 pp, medium confidence
5Gemini 4 Hydrogen48.6%estimated ± 1.7 pp, medium confidence
6Muse Wonder Full48.2%estimated ± 4.5 pp, medium confidence
7GPT-6.2 Astra46.0%estimated ± 4.8 pp, high confidence
8GPT-6.2 Sol45.7%estimated ± 4.8 pp, high confidence
9Gemini 4 Argon45.3%measured
10GPT-6 Astra44.3%measured
11Qwen4 Max44.2%estimated ± 4.5 pp, high confidence
12Ember-143.6%estimated ± 4.5 pp, high confidence
13GPT-6.1 Terra42.7%estimated ± 5.3 pp, low confidence
14Pareto 26.942.2%estimated ± 4.5 pp, high confidence
15GPT-6.1 Sol41.9%estimated ± 2.2 pp, high confidence
16MAI Code 241.3%estimated ± 4.5 pp, high confidence
17SWE-241.1%estimated ± 4.5 pp, high confidence
18Particle-140.6%estimated ± 4.5 pp, high confidence
19Claude Fable 5.140.2%measured
20Claude Mythos 540.2%estimated ± 4.4 pp, high confidence
21MiMo-V2.6-Pro39.9%estimated ± 4.5 pp, high confidence
22Kimi K3.1 Pro39.5%estimated ± 2.2 pp, high confidence
23MiniMax M3.1 Flash39.3%measured
24ERNIE 5.5 Flash Preview39.2%estimated ± 4.5 pp, high confidence
25Hy438.1%estimated ± 2.2 pp, high confidence
26Pareto 26.10 Preview37.8%estimated ± 4.5 pp, high confidence
27GPT-6 Sol36.7%estimated ± 4.5 pp, high confidence
28Muse Wonder36.5%estimated ± 1.7 pp, medium confidence
29GPT-5.6 Sol36.2%measured
30MiMo-V2.6-Flash35.8%estimated ± 4.5 pp, high confidence
31Claude Haiku 5.535.7%estimated ± 1.7 pp, medium confidence
32DeepSeek V4.1 Flash 101735.7%estimated ± 1.7 pp, medium confidence
33Grok 4.735.6%estimated ± 1.7 pp, medium confidence
34GLM-5.335.6%estimated ± 1.7 pp, medium confidence
35Muse Spark 1.335.6%estimated ± 1.7 pp, low confidence
36Grok 4.635.6%estimated ± 1.7 pp, low confidence
37GPT-5.6 Terra35.6%estimated ± 1.7 pp, low confidence
38GLM-5.3-Flash35.6%estimated ± 1.7 pp, low confidence
39GPT-5.6 Luna35.6%estimated ± 1.7 pp, low confidence
40Claude Sonnet 535.6%estimated ± 1.7 pp, low confidence
41Composer 2.535.6%estimated ± 1.7 pp, low confidence
42Step 5 Preview35.6%estimated ± 4.5 pp, high confidence
43DeepSeek V4 Pro35.1%estimated ± 3.7 pp, high confidence
44Claude Opus 535.0%measured
45Sakana Fugu-Ultra34.9%estimated ± 4.4 pp, high confidence
46Inkling 2 Small34.6%estimated ± 2.2 pp, high confidence
47GPT-6 Luna34.6%estimated ± 4.5 pp, high confidence
48Claude Fable 534.4%estimated ± 3.7 pp, high confidence
49Claude Opus 4.832.9%measured
50Hy4 flash32.7%estimated ± 4.5 pp, high confidence
51LongCat-2.532.7%estimated ± 4.5 pp, high confidence
52Mistral Large 432.6%estimated ± 4.8 pp, high confidence
53Kimi K332.0%measured
54GLM-5.231.7%measured
55Ling 3.1 Flash31.4%estimated ± 4.4 pp, high confidence
56DeepSeek V4 Flash 073130.8%estimated ± 3.7 pp, high confidence
57SWE-1.730.3%estimated ± 4.8 pp, high confidence
58Ultramarine 1 Flash30.0%estimated ± 4.5 pp, high confidence
59Beam 1.0 501B-A23B29.6%estimated ± 4.4 pp, high confidence
60Beam 1.0 501B-A23B FP829.4%estimated ± 4.4 pp, high confidence
61Muse Spark 1.229.3%estimated ± 3.7 pp, high confidence
62Hy4 preview29.3%estimated ± 4.4 pp, high confidence
63Beam 1.0 501B-A23B NVFP429.2%estimated ± 4.4 pp, high confidence
64Ornith-1.5-397B28.9%estimated ± 4.4 pp, high confidence
65Qwen3.8-27B28.9%estimated ± 3.7 pp, high confidence
66SWE-2 Flash28.8%estimated ± 4.5 pp, high confidence
67Muse Spark Open28.7%estimated ± 4.5 pp, high confidence
68MAI Code 2 Flash28.7%estimated ± 4.5 pp, high confidence
69Qwen3.8 Max28.6%estimated ± 3.7 pp, high confidence
70Claude Opus 4.728.6%measured
71Solar Pro 428.0%estimated ± 7.4 pp, low confidence
72Qwen3.8-Omni-Flash27.7%estimated ± 4.4 pp, high confidence
73Gemini 2.5 Pro27.3%estimated ± 7.4 pp, low confidence
74GPT-5.527.2%measured
75Qwen3.8-Flash-Next27.2%estimated ± 4.4 pp, high confidence
76Ornith-1.0-397B27.0%estimated ± 4.4 pp, high confidence
77Qwen3.8 Max Preview27.0%estimated ± 5.7 pp, medium confidence
78Gemma 4 31B27.0%estimated ± 7.4 pp, low confidence
79Ling 3.0 Flash VL26.6%estimated ± 7.4 pp, low confidence
80Ripple-65B Preview26.5%estimated ± 4.4 pp, high confidence
81Sage-A1.526.4%estimated ± 4.4 pp, high confidence
82dots3-note Preview26.3%estimated ± 4.4 pp, high confidence
83Inkling-Small26.0%estimated ± 3.7 pp, high confidence
84K-EXAONE 2.025.9%estimated ± 7.4 pp, low confidence
85Muse Spark 1.125.8%estimated ± 3.7 pp, high confidence
86A.X K225.7%estimated ± 7.4 pp, low confidence
87Trinity-Large-Preview25.6%estimated ± 7.4 pp, low confidence
88Trinity-Large-Thinking25.6%estimated ± 7.4 pp, low confidence
89Nemotron 3.5 Bolt 74B-A9B25.5%estimated ± 7.4 pp, low confidence
90Qwen3.5-122B-A10B25.5%estimated ± 7.4 pp, low confidence
91Atria Dawn Preview25.4%estimated ± 4.4 pp, high confidence
92Ornith-1.5-35B-A3B25.4%estimated ± 4.4 pp, high confidence
93DeepSeek V3 032425.3%estimated ± 7.4 pp, low confidence
94GPT-OSS 20B25.3%estimated ± 7.4 pp, low confidence
95Laguna S 2.125.3%estimated ± 4.4 pp, high confidence
96Mistral Small 425.3%estimated ± 7.4 pp, low confidence
97North Mini Code25.3%estimated ± 7.4 pp, low confidence
98Command A+25.2%estimated ± 7.4 pp, low confidence
99Mercury 2.525.2%estimated ± 7.4 pp, low confidence
100Sakana Fugu25.1%estimated ± 4.4 pp, high confidence
101Mistral Large 325.0%estimated ± 7.4 pp, low confidence
102DeepSeek V324.9%estimated ± 7.4 pp, low confidence
103Gemini 3.7 Flash24.8%estimated ± 3.7 pp, high confidence
104Qwen3.7 Plus24.2%estimated ± 4.4 pp, medium confidence
105Claude Opus 4.624.2%estimated ± 1.3 pp, medium confidence
106Gemini 3.8 Flash24.1%estimated ± 3.7 pp, high confidence
107Gemini 3.6 Flash23.8%estimated ± 3.7 pp, high confidence
108Step 3.7 Flash23.5%estimated ± 4.4 pp, medium confidence
109Grok 4.523.5%measured
110Gemini 3.5 Flash23.1%estimated ± 3.7 pp, high confidence
111GLM-522.8%estimated ± 4.4 pp, medium confidence
112Claude Opus 4.6 (Adaptive)22.6%estimated ± 3.7 pp, high confidence
113Kimi K2.7 Code22.6%estimated ± 3.7 pp, high confidence
114Spark-X2.6-4B22.3%estimated ± 4.4 pp, medium confidence
115Inkling22.1%estimated ± 3.7 pp, medium confidence
116Gemini 3.1 Pro22.0%measured
117Claude Sonnet 4.621.9%estimated ± 3.7 pp, medium confidence
118MAI-Thinking-121.5%estimated ± 4.4 pp, medium confidence
119GPT-5.221.4%estimated ± 1.3 pp, medium confidence
120Claude Opus 4.5 Thinking21.0%estimated ± 3.7 pp, medium confidence
121GLM-5.121.0%estimated ± 3.7 pp, medium confidence
122Nemotron 3.5 Super VL 120B A12B20.9%estimated ± 4.4 pp, medium confidence
123Kimi K2.620.9%estimated ± 3.7 pp, medium confidence
124Nemotron 3.5 Super VL 120B A12B NVFP420.7%estimated ± 4.4 pp, medium confidence
125Muse Glimmer 30B20.6%estimated ± 4.4 pp, medium confidence
126Qwen3.5 397B20.4%estimated ± 4.4 pp, medium confidence
127Kimi K2.520.3%estimated ± 4.4 pp, medium confidence
128Ornith-1.0-35B20.2%estimated ± 4.4 pp, medium confidence
129Qwen3.5 Plus19.9%estimated ± 10.5 pp, low confidence
130MiniMax M2.519.9%estimated ± 10.5 pp, low confidence
131GPT-5 mini19.9%estimated ± 10.5 pp, low confidence
132GPT-5.1-Codex19.9%estimated ± 10.5 pp, low confidence
133Claude Haiku 4.5 Thinking19.9%estimated ± 10.5 pp, low confidence
134Qwen3 Max19.9%estimated ± 10.5 pp, low confidence
135GLM-4.619.9%estimated ± 10.5 pp, low confidence
136Grok 4.1 Fast (Reasoning)19.9%estimated ± 10.5 pp, low confidence
137Grok 4 Fast (Reasoning)19.9%estimated ± 10.5 pp, low confidence
138Gemini 3.5 Flash-Lite19.8%estimated ± 3.7 pp, medium confidence
139Gemini 3 Flash19.8%estimated ± 3.7 pp, medium confidence
140MiniMax M319.8%estimated ± 3.7 pp, medium confidence
141Qwen3.6-35B-A3B19.7%estimated ± 4.4 pp, medium confidence
142Laguna M.119.5%estimated ± 4.4 pp, medium confidence
143Muse Spark19.3%estimated ± 3.7 pp, medium confidence
144GPT-5.419.0%measured
145Mercury 318.9%estimated ± 4.5 pp, medium confidence
146MiMo-V2.5-Pro18.9%estimated ± 3.7 pp, medium confidence
147MiniMax M2.718.7%estimated ± 3.7 pp, medium confidence
148Laguna XS 2.118.7%estimated ± 4.4 pp, medium confidence
149Ornith-1.5-9B18.6%estimated ± 4.4 pp, medium confidence
150Qwen3.6 Plus18.4%estimated ± 3.7 pp, medium confidence
151GPT-5.1-Codex-Max18.2%estimated ± 1.3 pp, low confidence
152Laguna XS.218.0%estimated ± 4.4 pp, medium confidence
153GPT-5.4 mini18.0%estimated ± 3.7 pp, medium confidence
154Qwen 3.6 Max (preview)17.8%estimated ± 3.7 pp, medium confidence
155Grok 4.2017.3%estimated ± 3.7 pp, medium confidence
156Grok 4.316.6%estimated ± 3.7 pp, medium confidence
157Ornith-1.0-9B16.3%estimated ± 4.4 pp, medium confidence
158MiMo-V2.516.2%estimated ± 3.7 pp, medium confidence
159Quasar 438B16.2%estimated ± 5.7 pp, low confidence
160Apodex 1.115.9%estimated ± 5.7 pp, low confidence
161Apodex 1.1 Mini15.9%estimated ± 5.7 pp, low confidence
162Claude Sonnet 4.5 Thinking15.4%estimated ± 3.7 pp, medium confidence
163Qwen3.6-27B15.4%estimated ± 3.7 pp, medium confidence
164LongCat-Flash-Lite-Sparse15.3%estimated ± 4.4 pp, medium confidence
165GPT-5.115.2%estimated ± 3.7 pp, medium confidence
166GPT-5.4 nano15.2%estimated ± 3.7 pp, medium confidence
167Gemini 3 Pro15.0%estimated ± 1.3 pp, low confidence
168GLM-4.714.8%estimated ± 3.7 pp, medium confidence
169Hy314.6%estimated ± 5.7 pp, low confidence
170Hy3 Preview14.6%estimated ± 5.7 pp, low confidence
171GPT-514.5%estimated ± 3.7 pp, medium confidence
172Nemotron 3 Ultra14.5%estimated ± 3.7 pp, medium confidence
173Qwen3.7 Max14.3%estimated ± 3.7 pp, medium confidence
174GPT-5.3 Codex14.2%estimated ± 1.3 pp, low confidence
175GPT-5.2-Codex13.0%estimated ± 1.3 pp, low confidence
176Claude Opus 4.512.9%estimated ± 1.3 pp, low confidence
177Claude Haiku 4.512.5%estimated ± 3.7 pp, medium confidence
178Mistral Medium 3.5 128B12.3%estimated ± 3.7 pp, medium confidence
179Granite 4.2 30B12.0%estimated ± 4.4 pp, medium confidence
180Ling 3.0 Flash11.4%estimated ± 3.7 pp, medium confidence
181Gemma 4 26B A4B11.0%estimated ± 6.2 pp, low confidence
182Qwen3.5 Flash10.8%estimated ± 3.7 pp, medium confidence
183LLaDA2.2-flash10.6%estimated ± 4.4 pp, medium confidence
184Mellum2.1-12B-A2.5B-Thinking9.7%estimated ± 4.4 pp, medium confidence
185Gemini 3.1 Flash-Lite9.6%estimated ± 3.7 pp, medium confidence

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General