llmrun

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.

Models ranked by reasoning llmrun Composite Score
#Modelllmrun Composite ScoreCodingAgentsMathScienceReasoningBenchmarksVRAM
121DeepSeek V3.1open412 GB · 7 benchmarksCoding –Agents –Math –Science –Reasoning 1067412 GB
122Gemini 2.5 Pro Preview (Jun 05)8 benchmarksCoding –Agents –Math –Science 114Reasoning 1068–
123Qwen Plus (Apr 28, 2025)2 benchmarks—Coding –Agents –Math –Science –Reasoning 1062–
124Qwen3.6 27Bopen17.9 GB · 11 benchmarksCoding 102Agents –Math 114Science 111Reasoning 1051117.9 GB
125Gemini 2.5 Flash Preview (May 20)7 benchmarksCoding 100Agents –Math –Science –Reasoning 1047–
126Claude 3.7 Sonnet (Feb 19, 2025)10 benchmarksCoding –Agents –Math –Science 105Reasoning 10410–
127Gemini 2.5 Flash9 benchmarksCoding –Agents 94Math –Science –Reasoning 1049–
128Qwen3.6 35B A3Bopen22.0 GB · 11 benchmarksCoding 105Agents –Math 109Science 110Reasoning 1041122.0 GB
129Qwen3.6 Flash10 benchmarksCoding –Agents –Math 108Science –Reasoning 10310–
130Codex Mini (May 16, 2025)2 benchmarks—Coding –Agents –Math –Science –Reasoning 1022–
131O1 (Dec 17, 2024)13 benchmarksCoding 104Agents –Math 104Science 104Reasoning 10213–
132Gemma 4 26B A4B ITopen16.5 GB · 8 benchmarksCoding 104Agents –Math –Science 102Reasoning 102816.5 GB
133Mistral Medium 3.5 128Bopen82.8 GB · 9 benchmarksCoding 102Agents –Math –Science 101Reasoning 102982.8 GB
134GPT 5.4 2026 03.056 benchmarksCoding 118Agents –Math –Science –Reasoning 1016–
135Qwen3 235B A22Bopen143 GB · 9 benchmarksCoding 101Agents –Math –Science –Reasoning 1019143 GB
136O3 Mini (Jan 31, 2025)19 benchmarksCoding 103Agents –Math 106Science 106Reasoning 10019–
137DeepSeek R1 0528open412 GB · 10 benchmarksCoding 107Agents –Math –Science –Reasoning 10010412 GB
138O1 Preview (Sep 12, 2024)5 benchmarksCoding –Agents –Math –Science –Reasoning 995–
139Qwen3 235B A22B Instruct 2507open143 GB · 8 benchmarksCoding 102Agents –Math –Science –Reasoning 988143 GB
140Claude 3.5 Sonnet (Oct 22, 2024)10 benchmarksCoding 99Agents 100Math –Science 88Reasoning 9710–
141Grok 3 Mini3 benchmarksCoding –Agents –Math –Science –Reasoning 973–
142Qwen3.5 9Bopen6.6 GB · 8 benchmarksCoding –Agents –Math –Science 105Reasoning 9786.6 GB
143GPT OSS 120Bopen70.8 GB · 19 benchmarksCoding 98Agents 82Math –Science 103Reasoning 961970.8 GB
144GPT 4.5 Preview (Feb 27, 2025)10 benchmarksCoding 96Agents –Math –Science 98Reasoning 9610–
145GPT 5.4 Mini 2026 03.176 benchmarksCoding –Agents –Math 89Science –Reasoning 966–
146Mistral Small 4 119B 2603open81.6 GB · 3 benchmarks—Coding –Agents –Math –Science –Reasoning 96381.6 GB
147DeepSeek R1open412 GB · 12 benchmarksCoding 100Agents –Math –Science 99Reasoning 9512412 GB
148GPT 5.2 2025 12.116 benchmarksCoding –Agents –Math –Science –Reasoning 956–
149Qwen3 30B A3B Thinking 2507open19.4 GB · 7 benchmarksCoding –Agents –Math –Science 99Reasoning 94719.4 GB
150GPT 5.4 Nano 2026 03.176 benchmarksCoding –Agents –Math 95Science –Reasoning 946–
151Claude 3.5 Sonnet (Jun 20, 2024)6 benchmarksCoding –Agents –Math –Science –Reasoning 936–
152GPT 4 Turbo (Apr 09, 2024)6 benchmarksCoding –Agents –Math 75Science –Reasoning 936–
153Qwen3 30B A3B Instruct 2507open19.4 GB · 4 benchmarksCoding –Agents –Math –Science –Reasoning 93419.4 GB
154Kimi K2 Instructopen617 GB · 10 benchmarksCoding 102Agents 102Math –Science –Reasoning 9310617 GB
155GPT 4.1 (Apr 14, 2025)16 benchmarksCoding 99Agents –Math 92Science 96Reasoning 9216–
156Mistral Medium 25086 benchmarksCoding 92Agents –Math –Science 91Reasoning 926–
157Grok 34 benchmarksCoding –Agents –Math –Science –Reasoning 924–
158Qwen3 32Bopen22.6 GB · 8 benchmarksCoding 93Agents –Math –Science 96Reasoning 92822.6 GB
159GPT OSS 20Bopen13.1 GB · 10 benchmarksCoding 102Agents –Math –Science 93Reasoning 911013.1 GB
160GPT 4.1 Mini (Apr 14, 2025)15 benchmarksCoding 92Agents –Math 95Science 97Reasoning 9115–
161GPT 5.1 2025 11.137 benchmarksCoding –Agents –Math –Science 96Reasoning 917–
162GPT 4o (Nov 20, 2024)10 benchmarksCoding 82Agents –Math –Science 82Reasoning 9010–
163GPT 4o (May 13, 2024)6 benchmarksCoding –Agents –Math –Science –Reasoning 906–
164GPT 5 Nano (Aug 07, 2025)17 benchmarksCoding 102Agents –Math 108Science –Reasoning 9017–
165Magistral Medium 25062 benchmarks—Coding –Agents –Math –Science –Reasoning 892–
166GPT 4 (Jun 13)7 benchmarksCoding –Agents –Math –Science –Reasoning 897–
167Qwen3 14Bopen11.9 GB · 7 benchmarksCoding –Agents –Math –Science 94Reasoning 88711.9 GB
168Magistral Small 2506open17.8 GB · 4 benchmarksCoding –Agents –Math –Science –Reasoning 88417.8 GB
169Qwen2.5 72B Instructopen49.3 GB · 8 benchmarksCoding –Agents –Math –Science –Reasoning 88849.3 GB
170DeepSeek v3 0324open412 GB · 12 benchmarksCoding 100Agents –Math –Science 98Reasoning 8812412 GB
171Gemini 2.0 Flash 0016 benchmarksCoding –Agents –Math –Science –Reasoning 876–
172Grok 2 (Dec 12)5 benchmarksCoding –Agents –Math –Science –Reasoning 875–
173Gemini 2.5 Flash Lite Preview Thinking (Jun 17)5 benchmarksCoding 95Agents –Math –Science –Reasoning 875–
174Mistral Large Instruct 2411open79.8 GB · 5 benchmarksCoding –Agents –Math –Science –Reasoning 87579.8 GB
175DeepSeek v3open412 GB · 8 benchmarksCoding 97Agents –Math –Science 90Reasoning 878412 GB
176Mistral Large 3 675B Instruct 2512open454 GB · 6 benchmarksCoding 93Agents –Math –Science 97Reasoning 866454 GB
177Llama 3.1 405B Instructopen276 GB · 8 benchmarksCoding –Agents –Math –Science 85Reasoning 868276 GB
178DeepSeek Chat5 benchmarksCoding –Agents –Math –Science –Reasoning 865–
179Claude 3 Opus (Feb 29, 2024)9 benchmarksCoding –Agents –Math –Science –Reasoning 869–
180Mistral Large Instruct 2407open86.4 GB · 6 benchmarksCoding –Agents –Math –Science –Reasoning 86686.4 GB
181Llama 4 Maverick 17B 128E Instructopen273 GB · 15 benchmarksCoding 81Agents –Math –Science 96Reasoning 8615273 GB
182O1 Mini (Sep 12, 2024)7 benchmarksCoding 91Agents –Math –Science –Reasoning 857–
183Qwen3 30B A3Bopen19.4 GB · 6 benchmarksCoding –Agents –Math –Science –Reasoning 84619.4 GB
184GPT 4 Turbo2 benchmarks—Coding –Agents –Math –Science –Reasoning 842–
185C4ai Command A 03 2025open78.7 GB · 4 benchmarksCoding –Agents –Math –Science –Reasoning 84478.7 GB
186Llama 3.3 70B Instructopen51.3 GB · 12 benchmarksCoding 76Agents –Math –Science 81Reasoning 831251.3 GB
187Llama 3.1 70B Instructopen51.3 GB · 7 benchmarksCoding –Agents –Math –Science –Reasoning 83751.3 GB
188Gemini 1.5 Pro 0028 benchmarksCoding –Agents –Math –Science 89Reasoning 838–
189Qwen3 8Bopen7.6 GB · 7 benchmarksCoding –Agents –Math –Science 88Reasoning 8177.6 GB
190Llama 4 Scout 17B 16E Instructopen77.0 GB · 11 benchmarksCoding –Agents –Math –Science 84Reasoning 811177.0 GB
191Gemini 1.5 Pro 0014 benchmarksCoding –Agents –Math –Science –Reasoning 804–
192Mixtral 8x22B Instruct v0.1open88.4 GB · 5 benchmarksCoding –Agents –Math –Science –Reasoning 78588.4 GB
193GPT 4o (Aug 06, 2024)7 benchmarksCoding –Agents –Math 75Science –Reasoning 777–
194Gemma 3 27B ITopen21.7 GB · 9 benchmarksCoding 70Agents –Math –Science 81Reasoning 74921.7 GB
195C4ai Command R Plus 08 2024open73.8 GB · 3 benchmarks—Coding –Agents –Math –Science –Reasoning 74373.8 GB
196Gemini 1.5 Flash 0014 benchmarksCoding –Agents –Math –Science –Reasoning 744–
197Claude 2.15 benchmarksCoding –Agents –Math –Science –Reasoning 735–
198GPT 4o Mini (Jul 18, 2024)13 benchmarksCoding 68Agents –Math 76Science –Reasoning 7313–
199GPT 4.1 Nano (Apr 14, 2025)13 benchmarksCoding 75Agents –Math –Science 82Reasoning 7213–
200Mixtral 8x7B Instruct v0.1open30.5 GB · 4 benchmarksCoding –Agents –Math –Science –Reasoning 71430.5 GB
201Ministral 3B 24103 benchmarks—Coding –Agents –Math –Science –Reasoning 713–
202Llama 3.1 8B Instructopen7.9 GB · 9 benchmarksCoding 56Agents –Math –Science 56Reasoning 7097.9 GB
203Gemma 3 4B ITopen5.0 GB · 5 benchmarksCoding –Agents –Math –Science –Reasoning 6955.0 GB
204Claude 3 Haiku (Mar 07, 2024)6 benchmarksCoding –Agents –Math –Science –Reasoning 676–
205Gemma 3 12B ITopen10.9 GB · 7 benchmarksCoding –Agents –Math –Science 72Reasoning 66710.9 GB
206Gemma 2 27B2 benchmarks—Coding –Agents –Math –Science –Reasoning 662–
207Qwen2.5 7B Instructopen5.8 GB · 5 benchmarksCoding –Agents –Math –Science –Reasoning 6555.8 GB
208Meta Llama 3 8B Instructopen7.9 GB · 5 benchmarksCoding –Agents –Math –Science –Reasoning 6557.9 GB
209C4ai Command R 08 2024open25.1 GB · 2 benchmarks—Coding –Agents –Math –Science –Reasoning 65225.1 GB
210Llama 2 70B Chat HFopen50.2 GB · 4 benchmarksCoding –Agents –Math –Science –Reasoning 42450.2 GB
211GPT 3.5 Turbo (Jan 25)8 benchmarksCoding –Agents –Math 72Science –Reasoning 368–

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.