LLM Stats (CodeForces): leaderboard
Metric: Score. Source: llm-stats.com. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek V4 Pro (Max) | 100 |
| 2 | DeepSeek V4 Flash (Max) | 100 |
| 3 | DeepSeek V4 Flash (0423) | 93.87 |
| 4 | DeepSeek V3.2 Speciale | 90 |
| 5 | Qwen 3.5 122B A10B | 85.1 |
| 6 | Qwen 3.5 35B A3B | 82.2 |
| 7 | GPT-OSS-120B | 82.1 |
| 8 | Qwen 3.5 27B | 80.7 |
| 9 | DeepSeek V3.2 | 79.5 |
| 10 | DeepSeek V3.2 (Thinking) | 79.5 |
| 11 | GPT-OSS-20B | 74.33 |
| 12 | DeepSeek V3.2 Exp | 70.7 |
| 13 | DeepSeek V3.1 | 69.7 |
| 14 | Qwen 3 32B | 65.9 |
| 15 | DeepSeek R1 0528 | 64.33 |
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-codeforces · How It Works · Data refreshed daily, snapshot 2026-09-05.