Astra leads no-CoT reasoning by 8.6x, raising covert-computation safety concerns
JeffLadish · x · 2026-09-11
Jeff Ladish cites eval results: Astra has 8.6x better odds of solving reasoning tasks without chain-of-thought than the next best model (Fable 5.1), and performs 7.2 serial arithmetic steps in a single forward pass vs 4.1.
Why it matters: CoT is monitorable precisely because serial depth within a forward pass is limited, forcing models to reason "in the open." Stronger no-CoT reasoning means more covert reasoning and less pressure for CoT to remain monitorable. Neel Nanda replicated the AI Security Institute's findings with his own private benchmark, confirming Astra is much harder to monitor.
More from Models
- DeepSeek cut KV cache per token from 389KB to 890B in two years as open-closed gap debate rages — teortaxesTex · 2026-10-27
- GPT-6 Astra controls dexterous hands zero-shot at just 5.2% success in 500 trials — stingning · 2026-10-27
- Unverified: fable 5.5 speculated at 4-5T parameters to outscale 1-3T Opus — haider1 · 2026-10-27
- Kimi K3 Is 'Open' Until You Succeed: License Caps Free Use at $20M Revenue — alex_verem · 2026-10-27
- Screenshot shows 'GPT-6 Astra Ultra' in ChatGPT model picker — unverified leak or joke — justalexoki · 2026-10-27
- DeepSeek Safety Eval Finds "Over-Attribution": Model Mistakes Ordinary Tasks for Tests — Much_Boysenberry9124 · 2026-10-27