Edit2TikZ - Compilation Success Rate: leaderboard
Metric: Compiled and rendered outputs (%). Source: arxiv.org. 14 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 88.37 |
| 2 | Claude Opus 4.8 | 78.88 |
| 3 | GPT-5.6 Terra | 77.33 |
| 4 | GPT-5.6 Sol | 75.65 |
| 5 | Claude Sonnet 4.6 | 74.1 |
| 6 | GPT-5.6 Luna | 74.1 |
| 7 | Qwen 3.6 27B | 69.77 |
| 8 | Qwen 3.7 Plus | 67.05 |
| 9 | Seed 2.1 Pro | 65.76 |
| 10 | gemma-4-E4B-it | 52.39 |
| 11 | Qwen 3.5 9B | 52.26 |
| 12 | Qwen 3.5 4B | 45.35 |
| 13 | Qwen 3.5 2B | 13.24 |
| 14 | Gemma 4 E2B (IT) | 6.72 |
Interactive version: theaggregate.ai/benchmark?slug=edit2tikz-compilation-success-rate · How It Works · Data refreshed daily, snapshot 2026-09-19.