Epoch AI - Algotune: leaderboard
Epoch AI AlgoTune evaluates coding agents on optimizing algorithm implementations for speed while preserving correctness.
Metric: Score. Source: epoch.ai. Saturation forecast: Around July 2027. 18 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.2 (2025-12-11) (Medium) | 2.05 | |
| 2 | Gemini 3.1 Pro (Preview) | 2.02 | #58 |
| 3 | GPT-5.4 (2026-03-05) (High) | 1.85 | #59 (GPT-5.4 (2026-03-05)) |
| 4 | Gemini 3 Pro (Preview) | 1.83 | #66 |
| 5 | O4 Mini (2025-04-16) (High) | 1.72 | #179 (O4 Mini (2025-04-16)) |
| 6 | DeepSeek R1 | 1.7 | #251 |
| 7 | GPT-5 (2025-08-07) (High) | 1.67 | |
| 8 | GLM-4.5 (Thinking) | 1.52 | #275 (GLM-4.5) |
| 9 | Gemini 2.5 Pro | 1.51 | #152 |
| 10 | Claude Opus 4.6 (High) | 1.47 | #63 (Claude Opus 4.6) |
| 11 | Qwen 3 Coder 480B A35B Instruct | 1.44 | #315 |
| 12 | GPT-OSS-120B (High) | 1.41 | #341 (GPT-OSS-120B) |
| 13 | GPT-5 Mini (2025-08-07) (High) | 1.38 | #172 (GPT-5 Mini (2025-08-07)) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=epoch-ai-algotune · How It Works · Data refreshed daily, snapshot 2026-10-20.