Verilog50 — Syntax validity — final iterative workflow: leaderboard
Table III final per-model workflow: 50 RTL tasks, 3 repetitions, iterative generation/repair with at most 5 cycles; syntax and functional simulation are different metrics. Source README explicitly identifies Llama 3 8B Instruct; TinyLlama Ollama tag does not establish a precise chat/pretraining checkpoint. Do not alias its source-qualified identity. Synthesis cells are not selected because README says synthesis was not computed. Metric: Syntax validity (%). Higher is better. Source: https://arxiv.org/html/2607.22759v1
Metric: Syntax validity (%). Source: arxiv.org. Saturation forecast: Around January 2027. 3 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 3 8B Instruct | 72 | #1119 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #7 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=verilog50-syntax-validity-final-iterative-workflow · How It Works · Data refreshed daily, snapshot 2026-10-17.