Fable 5 fails on basic physics Olympiad problem despite SOTA reasoning gains
geoffwolfe · x · 2026-08-17
Despite the strong reasoning capabilities of SOTA LLMs, they still struggle with basic physics questions. Fable 5 High immediately made a wrong assumption on problem 1 of the 2025 Online Physics Olympiad and solved it incorrectly, highlighting persistent gaps in rigorous reasoning.
More from Models
- Qwen3.8-27B-AEON-ULTIMATE-UNCENSORED released with BF16 weights — AIandDesign · 2026-08-17
- Qwen3.8-27B-Ridge-3.7bpw released, shrinking model size to 11.7GB — udmrzn · 2026-08-17
- Google open-sources TimesFM: A time-series foundation model trained on 100B+ data points — udmrzn · 2026-08-17
- Qwen 3.8 27B review: excellent model defaults to hilarious overthinking — marlene_zw · 2026-08-17
- Claude down again, Grok running reliably across the board — XFreeze · 2026-08-17
- Opinion: Anthropic Model 2 gains underwhelming, open source gap narrows — ricklamers · 2026-08-17