EpochAI Updates Leaderboard: Claude Fable 5 Beats GPT-5.6 in Long-Horizon Tasks
Jsevillamol · x · 2026-08-04
EpochAI has updated its MirrorCode long-horizon benchmark leaderboard. Claude Fable 5 leads the board with a 64% solve rate, significantly outperforming GPT-5.6 Sol, which achieved a 20% solve rate.
Related event: Claude Fable 5 Tops EpochAI MirrorCode Leaderboard(3 posts)→
More from Models
- Alibaba’s Qwen3.8-Max is set to open its 2.4T weights next week — ivan_bezdomny · 2026-08-04
- Databricks Tops Kimi K3 Inference Speed at 239 tokens/s — altryne · 2026-08-04
- Kimi is praised for long-horizon coding and unusually aligned behavior — casper_hansen_ · 2026-08-04
- llama.cpp patches boost DeepSeek-V4-Flash-0731 from 3.26 to 25.91 tok/s — dyn___ · 2026-08-04
- Qwen3.8-Max ranks No. 2 in Vision Arena, 13 points behind Claude Fable 5 — arena · 2026-08-04
- Polymarket puts an 82% chance on a new Google Gemini Pro model within two weeks — Polymarket · 2026-08-04