Agent Plasticity metric: GPT-5.6 Sol learns ~5x more efficiently than Claude Fable 5
anirudhg9119 · x · 2026-10-09
The author introduces Agent Plasticity, a new metric for agent learning ability:
- Definition: held-out performance gain per unit of learning cost, measured up to saturation.
- Across Chess, Go, and Hex, Claude Fable 5 reaches the highest fitted performance, but GPT-5.6 Sol learns 5x more efficiently.
- Core claim: capability is not the same as the ability to learn — evaluating agents requires measuring improvement, not just current skill.
More from Research
- Study finds female AI agents 'paid' less than male ones, extending gender bias to agents — TMWNN · 2026-10-09
- LightOnOCR-3 training: using OCR reference text to guide logical block grouping — IgorCarron · 2026-10-09
- No single layout model works everywhere: LightOnOCR-3 team on combining complementary sources — IgorCarron · 2026-10-09
- Text2Sim turns text prompts into editable physics-grounded videos and sim setups — erwincoumans · 2026-10-09
- Building Rome from a single image: new method reconstructs 3D scenes with geometry beyond the visible view — jonstephens85 · 2026-10-09
- PredActor unifies humanoid motion generation and control in one diffusion policy at 50Hz on Unitree G1 — carlosdponx · 2026-10-09