Coding Agent Leaderboard - Shellbench: leaderboard
Metric: Score (%). Source: huggingface.co. 7 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GLM-5.2-FP8 + OpenClaw | 18.7 |
| 2 | glm-5.2 + OpenClaw | 18.3 |
| 3 | Qwen3.6-27B-FP8 + OpenClaw | 12.8 |
| 4 | NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4 + OpenClaw | 1.4 |
| 5 | Mistral-Small-4-119B-2603-NVFP4 + OpenClaw | 0.9 |
| 6 | gemma-4-31B-it-FP8-block + OpenClaw | 0 |
| 7 | gpt-oss-120b + OpenClaw | 0 |
Interactive version: theaggregate.ai/benchmark?slug=coding-agent-leaderboard-shellbench · How It Works · Data refreshed daily, snapshot 2026-09-21.