llmrun Composite Score · Reasoning
The best LLMs for reasoning, open and proprietary together, on one open-ended scale where 100 is the average model (reasoning). Switch to open models to see only what you can download and run.
46 benchmarks4 sourcesUpdated 14 Nov 2026Reference scale 2026-Q4How the Composite Score is built
Reasoning score vs VRAM needed
This chart doesn't show proprietary models.
| # | Model | llmrun Composite Score | Coding | Agents | Math | Science | Reasoning | Benchmarks | VRAM |
|---|---|---|---|---|---|---|---|---|---|
| 121 | DeepSeek V3.1open412 GB · 7 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 106 | 7 | 412 GB | |
| 122 | Gemini 2.5 Pro Preview (Jun 05)8 benchmarks | Coding – | Agents – | Math – | Science 114 | Reasoning 106 | 8 | – | |
| 123 | Qwen Plus (Apr 28, 2025)2 benchmarks | — | Coding – | Agents – | Math – | Science – | Reasoning 106 | 2 | – |
| 124 | Qwen3.6 27Bopen17.9 GB · 11 benchmarks | Coding 102 | Agents – | Math 114 | Science 111 | Reasoning 105 | 11 | 17.9 GB | |
| 125 | Gemini 2.5 Flash Preview (May 20)7 benchmarks | Coding 100 | Agents – | Math – | Science – | Reasoning 104 | 7 | – | |
| 126 | Claude 3.7 Sonnet (Feb 19, 2025)10 benchmarks | Coding – | Agents – | Math – | Science 105 | Reasoning 104 | 10 | – | |
| 127 | Gemini 2.5 Flash9 benchmarks | Coding – | Agents 94 | Math – | Science – | Reasoning 104 | 9 | – | |
| 128 | Qwen3.6 35B A3Bopen22.0 GB · 11 benchmarks | Coding 105 | Agents – | Math 109 | Science 110 | Reasoning 104 | 11 | 22.0 GB | |
| 129 | Qwen3.6 Flash10 benchmarks | Coding – | Agents – | Math 108 | Science – | Reasoning 103 | 10 | – | |
| 130 | Codex Mini (May 16, 2025)2 benchmarks | — | Coding – | Agents – | Math – | Science – | Reasoning 102 | 2 | – |
| 131 | O1 (Dec 17, 2024)13 benchmarks | Coding 104 | Agents – | Math 104 | Science 104 | Reasoning 102 | 13 | – | |
| 132 | Gemma 4 26B A4B ITopen16.5 GB · 8 benchmarks | Coding 104 | Agents – | Math – | Science 102 | Reasoning 102 | 8 | 16.5 GB | |
| 133 | Mistral Medium 3.5 128Bopen82.8 GB · 9 benchmarks | Coding 102 | Agents – | Math – | Science 101 | Reasoning 102 | 9 | 82.8 GB | |
| 134 | GPT 5.4 2026 03.056 benchmarks | Coding 118 | Agents – | Math – | Science – | Reasoning 101 | 6 | – | |
| 135 | Qwen3 235B A22Bopen143 GB · 9 benchmarks | Coding 101 | Agents – | Math – | Science – | Reasoning 101 | 9 | 143 GB | |
| 136 | O3 Mini (Jan 31, 2025)19 benchmarks | Coding 103 | Agents – | Math 106 | Science 106 | Reasoning 100 | 19 | – | |
| 137 | DeepSeek R1 0528open412 GB · 10 benchmarks | Coding 107 | Agents – | Math – | Science – | Reasoning 100 | 10 | 412 GB | |
| 138 | O1 Preview (Sep 12, 2024)5 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 99 | 5 | – | |
| 139 | Qwen3 235B A22B Instruct 2507open143 GB · 8 benchmarks | Coding 102 | Agents – | Math – | Science – | Reasoning 98 | 8 | 143 GB | |
| 140 | Claude 3.5 Sonnet (Oct 22, 2024)10 benchmarks | Coding 99 | Agents 100 | Math – | Science 88 | Reasoning 97 | 10 | – | |
| 141 | Grok 3 Mini3 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 97 | 3 | – | |
| 142 | Qwen3.5 9Bopen6.6 GB · 8 benchmarks | Coding – | Agents – | Math – | Science 105 | Reasoning 97 | 8 | 6.6 GB | |
| 143 | GPT OSS 120Bopen70.8 GB · 19 benchmarks | Coding 98 | Agents 82 | Math – | Science 103 | Reasoning 96 | 19 | 70.8 GB | |
| 144 | GPT 4.5 Preview (Feb 27, 2025)10 benchmarks | Coding 96 | Agents – | Math – | Science 98 | Reasoning 96 | 10 | – | |
| 145 | GPT 5.4 Mini 2026 03.176 benchmarks | Coding – | Agents – | Math 89 | Science – | Reasoning 96 | 6 | – | |
| 146 | Mistral Small 4 119B 2603open81.6 GB · 3 benchmarks | — | Coding – | Agents – | Math – | Science – | Reasoning 96 | 3 | 81.6 GB |
| 147 | DeepSeek R1open412 GB · 12 benchmarks | Coding 100 | Agents – | Math – | Science 99 | Reasoning 95 | 12 | 412 GB | |
| 148 | GPT 5.2 2025 12.116 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 95 | 6 | – | |
| 149 | Qwen3 30B A3B Thinking 2507open19.4 GB · 7 benchmarks | Coding – | Agents – | Math – | Science 99 | Reasoning 94 | 7 | 19.4 GB | |
| 150 | GPT 5.4 Nano 2026 03.176 benchmarks | Coding – | Agents – | Math 95 | Science – | Reasoning 94 | 6 | – | |
| 151 | Claude 3.5 Sonnet (Jun 20, 2024)6 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 93 | 6 | – | |
| 152 | GPT 4 Turbo (Apr 09, 2024)6 benchmarks | Coding – | Agents – | Math 75 | Science – | Reasoning 93 | 6 | – | |
| 153 | Qwen3 30B A3B Instruct 2507open19.4 GB · 4 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 93 | 4 | 19.4 GB | |
| 154 | Kimi K2 Instructopen617 GB · 10 benchmarks | Coding 102 | Agents 102 | Math – | Science – | Reasoning 93 | 10 | 617 GB | |
| 155 | GPT 4.1 (Apr 14, 2025)16 benchmarks | Coding 99 | Agents – | Math 92 | Science 96 | Reasoning 92 | 16 | – | |
| 156 | Mistral Medium 25086 benchmarks | Coding 92 | Agents – | Math – | Science 91 | Reasoning 92 | 6 | – | |
| 157 | Grok 34 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 92 | 4 | – | |
| 158 | Qwen3 32Bopen22.6 GB · 8 benchmarks | Coding 93 | Agents – | Math – | Science 96 | Reasoning 92 | 8 | 22.6 GB | |
| 159 | GPT OSS 20Bopen13.1 GB · 10 benchmarks | Coding 102 | Agents – | Math – | Science 93 | Reasoning 91 | 10 | 13.1 GB | |
| 160 | GPT 4.1 Mini (Apr 14, 2025)15 benchmarks | Coding 92 | Agents – | Math 95 | Science 97 | Reasoning 91 | 15 | – | |
| 161 | GPT 5.1 2025 11.137 benchmarks | Coding – | Agents – | Math – | Science 96 | Reasoning 91 | 7 | – | |
| 162 | GPT 4o (Nov 20, 2024)10 benchmarks | Coding 82 | Agents – | Math – | Science 82 | Reasoning 90 | 10 | – | |
| 163 | GPT 4o (May 13, 2024)6 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 90 | 6 | – | |
| 164 | GPT 5 Nano (Aug 07, 2025)17 benchmarks | Coding 102 | Agents – | Math 108 | Science – | Reasoning 90 | 17 | – | |
| 165 | Magistral Medium 25062 benchmarks | — | Coding – | Agents – | Math – | Science – | Reasoning 89 | 2 | – |
| 166 | GPT 4 (Jun 13)7 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 89 | 7 | – | |
| 167 | Qwen3 14Bopen11.9 GB · 7 benchmarks | Coding – | Agents – | Math – | Science 94 | Reasoning 88 | 7 | 11.9 GB | |
| 168 | Magistral Small 2506open17.8 GB · 4 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 88 | 4 | 17.8 GB | |
| 169 | Qwen2.5 72B Instructopen49.3 GB · 8 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 88 | 8 | 49.3 GB | |
| 170 | DeepSeek v3 0324open412 GB · 12 benchmarks | Coding 100 | Agents – | Math – | Science 98 | Reasoning 88 | 12 | 412 GB | |
| 171 | Gemini 2.0 Flash 0016 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 87 | 6 | – | |
| 172 | Grok 2 (Dec 12)5 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 87 | 5 | – | |
| 173 | Gemini 2.5 Flash Lite Preview Thinking (Jun 17)5 benchmarks | Coding 95 | Agents – | Math – | Science – | Reasoning 87 | 5 | – | |
| 174 | Mistral Large Instruct 2411open79.8 GB · 5 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 87 | 5 | 79.8 GB | |
| 175 | DeepSeek v3open412 GB · 8 benchmarks | Coding 97 | Agents – | Math – | Science 90 | Reasoning 87 | 8 | 412 GB | |
| 176 | Mistral Large 3 675B Instruct 2512open454 GB · 6 benchmarks | Coding 93 | Agents – | Math – | Science 97 | Reasoning 86 | 6 | 454 GB | |
| 177 | Llama 3.1 405B Instructopen276 GB · 8 benchmarks | Coding – | Agents – | Math – | Science 85 | Reasoning 86 | 8 | 276 GB | |
| 178 | DeepSeek Chat5 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 86 | 5 | – | |
| 179 | Claude 3 Opus (Feb 29, 2024)9 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 86 | 9 | – | |
| 180 | Mistral Large Instruct 2407open86.4 GB · 6 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 86 | 6 | 86.4 GB | |
| 181 | Llama 4 Maverick 17B 128E Instructopen273 GB · 15 benchmarks | Coding 81 | Agents – | Math – | Science 96 | Reasoning 86 | 15 | 273 GB | |
| 182 | O1 Mini (Sep 12, 2024)7 benchmarks | Coding 91 | Agents – | Math – | Science – | Reasoning 85 | 7 | – | |
| 183 | Qwen3 30B A3Bopen19.4 GB · 6 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 84 | 6 | 19.4 GB | |
| 184 | GPT 4 Turbo2 benchmarks | — | Coding – | Agents – | Math – | Science – | Reasoning 84 | 2 | – |
| 185 | C4ai Command A 03 2025open78.7 GB · 4 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 84 | 4 | 78.7 GB | |
| 186 | Llama 3.3 70B Instructopen51.3 GB · 12 benchmarks | Coding 76 | Agents – | Math – | Science 81 | Reasoning 83 | 12 | 51.3 GB | |
| 187 | Llama 3.1 70B Instructopen51.3 GB · 7 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 83 | 7 | 51.3 GB | |
| 188 | Gemini 1.5 Pro 0028 benchmarks | Coding – | Agents – | Math – | Science 89 | Reasoning 83 | 8 | – | |
| 189 | Qwen3 8Bopen7.6 GB · 7 benchmarks | Coding – | Agents – | Math – | Science 88 | Reasoning 81 | 7 | 7.6 GB | |
| 190 | Llama 4 Scout 17B 16E Instructopen77.0 GB · 11 benchmarks | Coding – | Agents – | Math – | Science 84 | Reasoning 81 | 11 | 77.0 GB | |
| 191 | Gemini 1.5 Pro 0014 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 80 | 4 | – | |
| 192 | Mixtral 8x22B Instruct v0.1open88.4 GB · 5 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 78 | 5 | 88.4 GB | |
| 193 | GPT 4o (Aug 06, 2024)7 benchmarks | Coding – | Agents – | Math 75 | Science – | Reasoning 77 | 7 | – | |
| 194 | Gemma 3 27B ITopen21.7 GB · 9 benchmarks | Coding 70 | Agents – | Math – | Science 81 | Reasoning 74 | 9 | 21.7 GB | |
| 195 | C4ai Command R Plus 08 2024open73.8 GB · 3 benchmarks | — | Coding – | Agents – | Math – | Science – | Reasoning 74 | 3 | 73.8 GB |
| 196 | Gemini 1.5 Flash 0014 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 74 | 4 | – | |
| 197 | Claude 2.15 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 73 | 5 | – | |
| 198 | GPT 4o Mini (Jul 18, 2024)13 benchmarks | Coding 68 | Agents – | Math 76 | Science – | Reasoning 73 | 13 | – | |
| 199 | GPT 4.1 Nano (Apr 14, 2025)13 benchmarks | Coding 75 | Agents – | Math – | Science 82 | Reasoning 72 | 13 | – | |
| 200 | Mixtral 8x7B Instruct v0.1open30.5 GB · 4 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 71 | 4 | 30.5 GB | |
| 201 | Ministral 3B 24103 benchmarks | — | Coding – | Agents – | Math – | Science – | Reasoning 71 | 3 | – |
| 202 | Llama 3.1 8B Instructopen7.9 GB · 9 benchmarks | Coding 56 | Agents – | Math – | Science 56 | Reasoning 70 | 9 | 7.9 GB | |
| 203 | Gemma 3 4B ITopen5.0 GB · 5 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 69 | 5 | 5.0 GB | |
| 204 | Claude 3 Haiku (Mar 07, 2024)6 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 67 | 6 | – | |
| 205 | Gemma 3 12B ITopen10.9 GB · 7 benchmarks | Coding – | Agents – | Math – | Science 72 | Reasoning 66 | 7 | 10.9 GB | |
| 206 | Gemma 2 27B2 benchmarks | — | Coding – | Agents – | Math – | Science – | Reasoning 66 | 2 | – |
| 207 | Qwen2.5 7B Instructopen5.8 GB · 5 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 65 | 5 | 5.8 GB | |
| 208 | Meta Llama 3 8B Instructopen7.9 GB · 5 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 65 | 5 | 7.9 GB | |
| 209 | C4ai Command R 08 2024open25.1 GB · 2 benchmarks | — | Coding – | Agents – | Math – | Science – | Reasoning 65 | 2 | 25.1 GB |
| 210 | Llama 2 70B Chat HFopen50.2 GB · 4 benchmarks | Coding – | Agents – | Math – | Science – | Reasoning 42 | 4 | 50.2 GB | |
| 211 | GPT 3.5 Turbo (Jan 25)8 benchmarks | Coding – | Agents – | Math 72 | Science – | Reasoning 36 | 8 | – |
The range after each score is a 90% interval: given the boards a model has results on, its true score is very likely inside it. Short ranges mean many agreeing results; long ranges mean few or conflicting ones.
Why some models are missing. A model receives a score only if it has results on at least 4 benchmarks across at least 2 categories, with one being reasoning or coding, or on 3 benchmarks that narrow its 90% range to at most 15.6 points. The results must come from at least two different benchmarks: categories of one benchmark, such as LiveBench, count once. Skill scores require at least 2 benchmarks for that skill; otherwise they display as –.
VRAM is llmrun's estimate at Q4_K_M (or the smallest quantization we track) for the weights plus room for a 16K-token context. Longer contexts need more memory. Proprietary models can't run locally, so the VRAM filter hides them.
How the llmrun Composite Score is built · llmrun doesn't run these benchmarks; scores come from public sources.
Dataset. Grab the scores of every open model as a dataset on Hugging Face, including a timestamped snapshot for each refresh.