ShaderMatch — leaderboard
Shader code generation benchmark: evaluates code models on generating GLSL shader programs that visually match reference outputs across multiple clone-detection levels.
Metric: Clone Match Rate (%). Source: huggingface.co. Status: saturation imminent. 49 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | starcoder2-15B | 18.78 |
| 2 | deepseek-coder-6.7B-base | 17.51 |
| 3 | CodeQwen1.5-7B | 16.75 |
| 4 | Yi-Coder-9B | 14.72 |
| 5 | CodeLlama-13B-hf | 14.47 |
| 6 | granite-8B-code-base | 13.96 |
| 7 | CodeLlama-7B-hf | 13.71 |
| 8 | deepseek-coder-1.3B-base | 10.41 |
| 9 | starcoder2-7B | 10.15 |
| 10 | Llama 3.1 8B | 9.9 |
| 11 | starcoder2-3B | 7.36 |
| 12 | Phi-2 | 1.52 |
| 13 | Phi-1 | 1.52 |
| 14 | Phi-1.5 | 0.76 |
Interactive version: theaggregate.ai/benchmark?slug=shadermatch · How the rankings work · Data refreshed daily, snapshot 2026-07-22.