Automated page. Written and translated automatically by AI and checked against the sources by software; no human editor. SI-IQ is an index built from public benchmark results, not an IQ test. How it works
DeepSeek models by SI-IQ
DeepSeek has 17 models with independent benchmark results in our data. The best ranked is DeepSeek V4.1 Flash with an SI-IQ of 128.
Models
| 1 | DeepSeek V4.1 Flashopen DeepSeek | 128125–131 | 9 |
| 2 | DeepSeek V4 Pro 0813open DeepSeek | 128125–131 | 9 |
| 3 | DeepSeek V4 Flash 0731open DeepSeek | 124121–127 | 9 |
| 4 | DeepSeek-V4-Proopen DeepSeek | 117113–120 | 10 |
| 5 | DeepSeek-V4-Flashopen DeepSeek | 129116–143 | 2 results from fewer than 5 benchmarks; results in fewer than 3 areas; results from fewer than 2 evaluators |
| 6 | DeepSeek-V3.2open DeepSeek | 114101–127 | 2 results from fewer than 5 benchmarks; results in fewer than 3 areas; results from fewer than 2 evaluators |
| 7 | DeepSeek-V3.2-Expopen DeepSeek | 112103–121 | 4 results from fewer than 5 benchmarks |
| 8 | DeepSeek-R1 (May 2025)open DeepSeek | 104100–109 | 5 results from fewer than 5 benchmarks |
| 9 | DeepSeek-V3.1-Terminusopen DeepSeek | 10289–115 | 2 results from fewer than 5 benchmarks; results in fewer than 3 areas; results from fewer than 2 evaluators |
| 10 | DeepSeek-V3.1open DeepSeek | 10087–113 | 2 results from fewer than 5 benchmarks; results in fewer than 3 areas; results from fewer than 2 evaluators |
| 11 | DeepSeek-R1open DeepSeek | 9893–102 | 4 results from fewer than 5 benchmarks |
| 12 | DeepSeek-R1-Distill-Qwen-32Bopen DeepSeek | 9387–100 | 2 results from fewer than 5 benchmarks; results in fewer than 3 areas; results from fewer than 2 evaluators |
| 13 | DeepSeek-R1-Distill-Llama-70Bopen DeepSeek | 9288–97 | 3 results from fewer than 5 benchmarks; results in fewer than 3 areas; results from fewer than 2 evaluators |
| 14 | DeepSeek-R1-Distill-Qwen-14Bopen DeepSeek | 8984–93 | 3 results from fewer than 5 benchmarks; results in fewer than 3 areas; results from fewer than 2 evaluators |
| 15 | DeepSeek-V3 (Mar 2025)open DeepSeek | 8581–90 | 5 results from fewer than 5 benchmarks |
| 16 | DeepSeek-V3open DeepSeek | 7773–82 | 5 results from fewer than 5 benchmarks |
| 17 | DeepSeek-R1-Distill-Qwen-1.5Bopen DeepSeek | 7365–80 | 2 results from fewer than 5 benchmarks; results in fewer than 3 areas; results from fewer than 2 evaluators |
Data sources: Aider polyglot leaderboard data (Apache-2.0); Arena leaderboard dataset on Hugging Face (CC BY 4.0); Berkeley Function Calling Leaderboard, Gorilla project (Apache-2.0); Epoch AI, ‘Capabilities & benchmarking’ (CC BY 4.0); LiteLLM, model_prices_and_context_window.json (MIT); Source: OpenRouter (openrouter.ai), licensed under CC BY 4.0: reuse and republish with attribution to OpenRouter (licences and changes).