benchgap
Coding

PostTrainBench leaderboard

As of 2026-11-19, the highest measured score on PostTrainBench is 43.8% by Claude Opus 5.5. 132 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#ModelScoreSource
1Claude Opus 5.543.8%measured
2Gemini 4 Neon42.8%estimated ± 4.6 pp, high confidence
3Claude Fable 5.242.2%estimated ± 2.4 pp, medium confidence
4Claude Sonnet 5.542.2%estimated ± 2.4 pp, high confidence
5GPT-6.2 Astra42.0%estimated ± 4.6 pp, high confidence
6GPT-6 Astra41.9%measured
7GPT-6.2 Sol41.9%estimated ± 4.6 pp, high confidence
8Gemini 4 Carbon41.0%measured
9GPT-6.1 Terra40.7%estimated ± 3.3 pp, medium confidence
10Gemini 4 Argon40.4%measured
11GPT-5.6 Terra39.0%estimated ± 3.3 pp, medium confidence
12GPT-6.1 Sol38.8%measured
13DeepSeek V4.1 Flash 101738.4%estimated ± 3.3 pp, medium confidence
14Muse Wonder Full37.9%estimated ± 3.3 pp, medium confidence
15Gemini 4 Hydrogen37.5%estimated ± 2.4 pp, high confidence
16Claude Mythos 537.1%estimated ± 3.7 pp, high confidence
17Claude Fable 5.136.9%estimated ± 2.0 pp, high confidence
18GLM-5.3-Flash36.8%measured
19Kimi K3.1 Pro36.5%measured
20MiniMax M3.1 Flash36.1%estimated ± 2.0 pp, high confidence
21Claude Opus 535.7%measured
22GPT-6 Sol35.3%estimated ± 3.3 pp, medium confidence
23Hy435.2%measured
24SWE-235.0%estimated ± 4.3 pp, medium confidence
25DeepSeek V4 Pro34.8%estimated ± 3.7 pp, high confidence
26GLM-5.334.5%measured
27Claude Fable 533.2%estimated ± 2.4 pp, high confidence
28Claude Haiku 5.533.1%estimated ± 2.4 pp, high confidence
29Grok 4.733.1%estimated ± 2.4 pp, high confidence
30Muse Spark 1.333.1%estimated ± 2.4 pp, high confidence
31Grok 4.633.1%estimated ± 2.4 pp, high confidence
32Gemini 3.7 Flash33.1%estimated ± 2.4 pp, high confidence
33Gemini 3.8 Flash33.1%estimated ± 2.4 pp, high confidence
34Qwen3.8 Max33.1%estimated ± 2.4 pp, medium confidence
35Muse Spark 1.233.1%estimated ± 2.4 pp, medium confidence
36Inkling33.1%estimated ± 2.4 pp, medium confidence
37MiMo-V2.6-Pro32.5%estimated ± 5.3 pp, low confidence
38Sakana Fugu-Ultra32.2%estimated ± 3.7 pp, high confidence
39Inkling 2 Small31.9%measured
40Kimi K331.5%measured
41GPT-5.6 Sol31.0%measured
42Claude Opus 4.831.0%measured
43GPT-5.6 Luna31.0%estimated ± 3.3 pp, medium confidence
44Mistral Large 430.9%estimated ± 4.6 pp, high confidence
45Muse Spark Open30.6%estimated ± 4.6 pp, medium confidence
46GLM-5.230.4%measured
47DeepSeek V4 Flash 073129.3%estimated ± 3.7 pp, high confidence
48SWE-2 Flash29.3%estimated ± 4.3 pp, medium confidence
49Ling 3.1 Flash28.9%estimated ± 3.7 pp, high confidence
50Beam 1.0 501B-A23B27.3%estimated ± 3.7 pp, high confidence
51Beam 1.0 501B-A23B FP827.2%estimated ± 3.7 pp, high confidence
52Qwen3.8-27B27.0%estimated ± 3.7 pp, high confidence
53Hy4 preview27.0%estimated ± 3.7 pp, high confidence
54Beam 1.0 501B-A23B NVFP427.0%estimated ± 3.7 pp, high confidence
55Ornith-1.5-397B26.7%estimated ± 3.7 pp, high confidence
56SWE-1.725.7%estimated ± 4.3 pp, medium confidence
57Qwen3.8-Omni-Flash25.6%estimated ± 3.7 pp, high confidence
58Qwen3.8-Flash-Next25.1%estimated ± 3.7 pp, high confidence
59Qwen3.8 Max Preview25.0%estimated ± 5.9 pp, low confidence
60Ornith-1.0-397B24.9%estimated ± 3.7 pp, high confidence
61Ripple-65B Preview24.5%estimated ± 3.7 pp, high confidence
62Sage-A1.524.4%estimated ± 3.7 pp, high confidence
63Claude Opus 4.724.3%measured
64dots3-note Preview24.3%estimated ± 3.7 pp, high confidence
65Muse Wonder24.2%measured
66GPT-5.523.9%measured
67Inkling-Small23.7%estimated ± 3.7 pp, high confidence
68Muse Spark 1.123.5%estimated ± 3.7 pp, high confidence
69Atria Dawn Preview23.5%estimated ± 3.7 pp, high confidence
70Ornith-1.5-35B-A3B23.5%estimated ± 3.7 pp, high confidence
71Laguna S 2.123.3%estimated ± 3.7 pp, high confidence
72Sakana Fugu23.1%estimated ± 3.7 pp, high confidence
73Grok 4.522.6%measured
74Qwen3.7 Plus22.4%estimated ± 3.7 pp, medium confidence
75Step 3.7 Flash21.7%estimated ± 3.7 pp, medium confidence
76Claude Sonnet 521.4%estimated ± 3.7 pp, high confidence
77Composer 2.521.4%estimated ± 3.7 pp, high confidence
78Gemini 3.6 Flash21.4%estimated ± 3.7 pp, high confidence
79GLM-521.0%estimated ± 3.7 pp, medium confidence
80Gemini 3.5 Flash20.6%estimated ± 3.7 pp, high confidence
81Spark-X2.6-4B20.6%estimated ± 3.7 pp, medium confidence
82Claude Opus 4.620.4%estimated ± 1.2 pp, medium confidence
83Claude Opus 4.6 (Adaptive)20.1%estimated ± 3.7 pp, high confidence
84Kimi K2.7 Code20.1%estimated ± 3.7 pp, high confidence
85MAI-Thinking-119.8%estimated ± 3.7 pp, medium confidence
86Claude Sonnet 4.619.4%estimated ± 3.7 pp, medium confidence
87Nemotron 3.5 Super VL 120B A12B19.3%estimated ± 3.7 pp, medium confidence
88Nemotron 3.5 Super VL 120B A12B NVFP419.1%estimated ± 3.7 pp, medium confidence
89Muse Glimmer 30B19.0%estimated ± 3.7 pp, medium confidence
90Qwen3.5 397B18.9%estimated ± 3.7 pp, medium confidence
91GPT-5.218.8%estimated ± 1.2 pp, medium confidence
92Kimi K2.518.8%estimated ± 3.7 pp, medium confidence
93Ornith-1.0-35B18.6%estimated ± 3.7 pp, medium confidence
94Claude Opus 4.5 Thinking18.5%estimated ± 3.7 pp, medium confidence
95GLM-5.118.5%estimated ± 3.7 pp, medium confidence
96Kimi K2.618.3%estimated ± 3.7 pp, medium confidence
97Qwen3.6-35B-A3B18.2%estimated ± 3.7 pp, medium confidence
98Gemini 3.1 Pro18.1%measured
99Laguna M.118.0%estimated ± 3.7 pp, medium confidence
100GPT-5.417.8%measured
101Laguna XS 2.117.3%estimated ± 3.7 pp, medium confidence
102Ornith-1.5-9B17.2%estimated ± 3.7 pp, medium confidence
103GPT-5.1-Codex-Max17.2%estimated ± 1.2 pp, low confidence
104Gemini 3.5 Flash-Lite17.2%estimated ± 3.7 pp, medium confidence
105Gemini 3 Flash17.2%estimated ± 3.7 pp, medium confidence
106MiniMax M317.2%estimated ± 3.7 pp, medium confidence
107Qwen3.5 Plus17.0%estimated ± 8.3 pp, low confidence
108MiniMax M2.517.0%estimated ± 8.3 pp, low confidence
109GPT-5 mini17.0%estimated ± 8.3 pp, low confidence
110GPT-5.1-Codex17.0%estimated ± 8.3 pp, low confidence
111Claude Haiku 4.5 Thinking17.0%estimated ± 8.3 pp, low confidence
112Qwen3 Max17.0%estimated ± 8.3 pp, low confidence
113GLM-4.617.0%estimated ± 8.3 pp, low confidence
114Grok 4.1 Fast (Reasoning)17.0%estimated ± 8.3 pp, low confidence
115Gemini 2.5 Pro17.0%estimated ± 8.3 pp, low confidence
116Grok 4 Fast (Reasoning)17.0%estimated ± 8.3 pp, low confidence
117Muse Spark16.7%estimated ± 3.7 pp, medium confidence
118Laguna XS.216.6%estimated ± 3.7 pp, medium confidence
119MiMo-V2.5-Pro16.3%estimated ± 3.7 pp, medium confidence
120MiniMax M2.716.1%estimated ± 3.7 pp, medium confidence
121Qwen3.6 Plus15.8%estimated ± 3.7 pp, medium confidence
122Gemini 3 Pro15.7%estimated ± 1.2 pp, low confidence
123GPT-5.4 mini15.4%estimated ± 3.7 pp, medium confidence
124GPT-5.3 Codex15.4%estimated ± 1.2 pp, low confidence
125GPT-6 Luna15.3%estimated ± 3.3 pp, low confidence
126Qwen 3.6 Max (preview)15.3%estimated ± 3.7 pp, medium confidence
127Ornith-1.0-9B15.1%estimated ± 3.7 pp, medium confidence
128GPT-5.2-Codex14.9%estimated ± 1.2 pp, low confidence
129Claude Opus 4.514.8%estimated ± 1.2 pp, low confidence
130Grok 4.2014.7%estimated ± 3.7 pp, medium confidence
131LongCat-Flash-Lite-Sparse14.1%estimated ± 3.7 pp, medium confidence
132Grok 4.314.1%estimated ± 3.7 pp, medium confidence
133MiMo-V2.513.7%estimated ± 3.7 pp, medium confidence
134Claude Sonnet 4.5 Thinking12.9%estimated ± 3.7 pp, medium confidence
135Qwen3.6-27B12.9%estimated ± 3.7 pp, medium confidence
136GPT-5.112.7%estimated ± 3.7 pp, medium confidence
137GPT-5.4 nano12.7%estimated ± 3.7 pp, medium confidence
138GLM-4.712.4%estimated ± 3.7 pp, medium confidence
139Quasar 438B12.4%estimated ± 5.9 pp, low confidence
140GPT-512.1%estimated ± 3.7 pp, medium confidence
141Nemotron 3 Ultra12.1%estimated ± 3.7 pp, medium confidence
142Apodex 1.112.0%estimated ± 5.9 pp, low confidence
143Apodex 1.1 Mini12.0%estimated ± 5.9 pp, low confidence
144Qwen3.7 Max11.9%estimated ± 3.7 pp, medium confidence
145Granite 4.2 30B11.0%estimated ± 3.7 pp, medium confidence
146Claude Haiku 4.510.3%estimated ± 3.7 pp, medium confidence
147Mistral Medium 3.5 128B10.1%estimated ± 3.7 pp, medium confidence
148Gemma 4 26B A4B9.9%estimated ± 4.1 pp, low confidence
149LLaDA2.2-flash9.8%estimated ± 3.7 pp, medium confidence
150Ling 3.0 Flash9.3%estimated ± 3.7 pp, medium confidence
151Mellum2.1-12B-A2.5B-Thinking9.0%estimated ± 3.7 pp, medium confidence
152Qwen3.5 Flash8.7%estimated ± 3.7 pp, medium confidence
153Gemini 3.1 Flash-Lite7.7%estimated ± 3.7 pp, medium confidence

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General