Vision2Web S2-2026 (Claude Code) - Level 3 Full-Stack: leaderboard
Vision2Web Level 3: the agent develops a full-stack web application from requirements and prototypes; the score is the mean of a VLM-judged visual score and a functional score from workflow-driven GUI tests, on a 0-100 scale under the Claude Code framework, higher is better (He et al., 2026). Season S2-2026 (Apr to Jun 2026) uses Gemini-3.1-pro-preview as the visual judge and GLM-4.6V as the GUI testing agent; scores compare only within a season.
Metric: Score (0-100). Source: vision2web-bench.github.io. Saturation forecast: Around May 2027. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 37.3 | #100 |
| 2 | Claude Sonnet 4.6 | 35.3 | #158 |
| 3 | Claude Opus 4.5 | 31.9 | #152 |
| 4 | GPT-5.4 | 22.3 | #121 |
| 5 | Kimi K2.5 | 18.1 | #252 |
| 6 | Gemini 3.1 Pro (Preview) | 13.6 | #91 |
| 7 | Qwen 3.5 397B A17B (Thinking) | 5.2 | #242 (Qwen 3.5 397B A17B) |
| 8 | Seed 2.0 Pro | 3 | #209 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=vision2web-s2-2026-claude-code-level-3-full-stack · How It Works · Data refreshed daily, snapshot 2026-11-12.