AlgoTune
Speed up numerical algorithms; score is speedup.
Agent Index weight
0%
shown for context only
Models tested
18
from the benchmark’s leaderboard and reports
Results
| # | Model | Score | Normalised | Harness | Cost/task | Date |
|---|---|---|---|---|---|---|
| 1 | GPT-5.2 Supported OpenAI | 2.05× | source | — | — | |
| 2 | Gemini 3.1 Pro Preview Supported Google | 2.02× | source | — | — | |
| 3 | GPT-5.4 Supported OpenAI | 1.85× | source | — | — | |
| 4 | Gemini 3 Pro Preview Supported Google | 1.83× | source | — | — | |
| 5 | Claude Opus 4.5 (latest) Supported Anthropic | 1.77× | source | — | — | |
| 6 | o4-mini Estimated OpenAI | 1.72× | source | — | — | |
| 7 | DeepSeek-R1 Estimated DeepSeek | 1.70× | source | — | — | |
| 8 | GPT-5 Supported OpenAI | 1.67× | source | — | — | |
| 9 | Claude Sonnet 4.5 (latest) Supported Anthropic | 1.52× | source | — | — | |
| 10 | GLM-4.5 Z.ai (Zhipu) | 1.52× | source | — | — | |
| 11 | Gemini 2.5 Pro Supported Google | 1.51× | source | — | — | |
| 12 | Claude Opus 4.6 Supported Anthropic | 1.47× | source | — | — | |
| 13 | Qwen3 Coder 480B A35B Estimated Alibaba (Qwen) | 1.44× | source | — | — | |
| 14 | GPT OSS 120B Supported OpenAI | 1.41× | source | — | — | |
| 15 | GPT-5 Mini Estimated OpenAI | 1.38× | source | — | — | |
| 16 | Claude Opus 4.1 (latest) Supported Anthropic | 1.34× | source | — | — | |
| 17 | Claude-Opus-4 Supported Anthropic | 1.33× | source | — | — | |
| 18 | GPT-5 Pro OpenAI | 1.31× | source | — | — |