Opus 5 Tops Code Arena Image-to-WebDev, Beating GPT-5.6 and Grok-4.5
arena · x · 2026-08-02
Code Arena released the latest scores and rankings for its 'Image-to-WebDev' benchmark. The evaluation tests AI models on their ability to generate websites from images or screenshots, alongside agentic coding workflows involving multi-step reasoning and tool use.
Rankings:
- Opus 5 (Max): 1,669 pts
- GPT-5.6 Sol (xHigh): 1,581 pts
- Grok-4.5: 1,578 pts
- Kimi K3 (Max): 1,571 pts
- Muse Spark 1.1: 1,550 pts
- GPT-5.6 Terra: 1,531 pts
- GPT-5.6 Luna: 1,497 pts
Related event: Claude Opus 5 Tops AI Leaderboards for Web Development and Image-to-Code(4 posts)→
More from coding & agent
- Claude Builds Walkable 3D Jungle in Browser with Zero Downloaded Assets — prasenx · 2026-08-02
- AI Browsers Struggle as Users Shift to Firecrawl and Scraping Agents Over Google — EXM7777 · 2026-08-02
- Comp AI open-sources its agentic-first CRM under MIT license — JosephJacks_ · 2026-08-02
- Using Clairvoyance, Opus 5 coding agent consumes only 1% of weekly token limit — draginol · 2026-08-02
- Generating a Retro CRT Arcade Game in One Shot with Claude Opus — draginol · 2026-08-02
- Oxford & Tsinghua Launch Agent Memory Challenge for Unified LLM Memory Eval — JundeMorsenWu · 2026-08-02