Agent Arena Adds Categories: GPT 5.6 Sol Tops Code Ranking

arena · x · 2026-08-18

Agent Arena launched new categories with different top models: GPT 5.6 Sol (xHigh) for Code, Claude Opus 5 (High) for Work, and Claude Opus 5 (Max) for Chat.

The arena measures models on millions of real-world, long-horizon agentic tasks, where a single session can cost up to $1,000. They explain token pricing, caching, and the use of task-based cost estimates for leaderboards.

Related event: Agent Arena's New Categories: GPT 5.6 Tops Coding, Claude Opus Leads Work(2 posts)→

Original post →

More from coding & agent

coding & agent channel →