Wei-Lin Chiang
@infwinston
Co-founder & CTO @Arena, PhD in AI & systems @UCBerkeley
加入 February 2012
967 正在關注 5.7K 粉絲
Test Time Scaling in Agent Arena.
- Opus-5 test-time scaling beyond GPT-5.6 sol
- Opus-5-Max reaching Fable performance but more expensive
Claude Opus 5 vs GPT 5.6 in Agent Arena
Highlights:
- Opus (High) and Opus (Max) outperform GPT Sol (xHigh), but at higher cost.
- Opus (Medium) matches GPT Sol (xHigh) in performance at approximately the same cost.
- Fable 5 (High) achieves the most optimal price vs. performance trade off. Real-world cost reflects total task execution, not just per-token pricing, so additional iterations and tool calls increase overall expense.
顯示更多
