LLM2014 Code 2025-11: leaderboard
LLM2014 monthly coding benchmark for 2025-11, reporting the aggregate code-generation score across programming tasks.
Metric: Multi-round Score. Source: raw.githubusercontent.com. Saturation forecast: Estimated already saturated. 26 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro | 96 | #80 |
| 2 | GPT-5 Mini (High) | 93 | #186 (GPT-5 Mini) |
| 3 | Claude Opus 4.5 | 90.75 | #83 |
| 4 | Claude Sonnet 4.5 (Thinking) | 81.67 | #142 (Claude Sonnet 4.5) |
| 5 | Claude Sonnet 4.5 | 75.25 | #142 |
| 6 | Gemini 2.5 Pro | 74.36 | #152 |
| 7 | DeepSeek V3.2 (Thinking) | 68 | #205 (DeepSeek V3.2) |
| 8 | GLM-4.6 (Thinking) | 66.81 | #252 (GLM-4.6) |
| 9 | DeepSeek V3.2 | 65.28 | #205 |
| 10 | Qwen 3 Max | 63.95 | #208 |
| 11 | Kimi K2 (Thinking) | 55.75 | #243 (Kimi K2) |
| 12 | Kimi K2 0905 | 53.8 | #288 |
| 13 | MiniMax-M2 | 53.56 | #317 |
| 14 | Seed-1.6 (Thinking) | 53.56 | #265 (Seed-1.6) |
| 15 | Qwen 3 Coder 480B A35B | 52.22 | #279 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=llm2014-code-2025-11 · How It Works · Data refreshed daily, snapshot 2026-10-20.