TuRTLe Code Completion (Verilator): leaderboard
TuRTLe leaderboard variant for RTL code completion evaluated with Verilator across VerilogEval MC and VeriGen.
Metric: Average Score (%). Source: huggingface.co. Status: saturation imminent. 45 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Kimi K2.5 | 96.31 |
| 2 | DeepSeek R1 | 95.9 |
| 3 | DeepSeek V3.1 Terminus | 95.25 |
| 4 | DeepSeek R1 0528 | 94.97 |
| 5 | GLM-5 FP8 | 92.73 |
| 6 | GPT-OSS-120B | 91.28 |
| 7 | Gemma 4 31B (IT) | 90.94 |
| 8 | Qwen 2.5 32B | 89.99 |
| 9 | Llama 3.1 405B | 89.35 |
| 10 | starchat2-15B-v0.1 | 87.81 |
| 11 | GPT-OSS-20B | 86.53 |
| 12 | Qwen 2.5 72B | 84.24 |
| 13 | QwQ-32B | 80.47 |
| 14 | Hermes-4-14B | 74.06 |
| 15 | Qwen 3 8B | 72.79 |
Interactive version: theaggregate.ai/benchmark?slug=turtle-code-completion-verilator · How It Works · Data refreshed daily, snapshot 2026-09-05.