Observation: Astra uses filler tokens far more effectively than other models
scaling01 · x · 2026-09-26
scaling01, who has been tracking looping behavior, observes that Google's Astra appears to use filler tokens much more effectively than other models — it can leverage not only user-provided filler tokens but also filler tokens it generates itself. The observation hints at differentiated internal computation strategies and warrants further verification.
More from Models
- Interfaze AI Open-Sources Lev, a 4B System-One Model Built on Qwen — ycombinator · 2026-09-26
- Hugo Bowne-Anderson: claims that Opus 5.5 writes better prose are nonsense — hugobowne · 2026-09-26
- GPT-6 Sol Tops Vercel's DeepSecBench With Highest Score at 20% of Runner-Up's Cost — cramforce · 2026-09-26
- Alibaba Claims RSI Progress: Qwen3.8-Max Self-Improved Its AA Score From 40 to 45 Over 33 Autonomous Cycles — teortaxesTex · 2026-09-26
- Opus 5.5 unblocks a 4-month ts-rust port in 10 hours where GPT-6 Astra stalled at 85% — EricBuess · 2026-09-26
- Kev-4B is now available on OpenRouter — charles_irl · 2026-09-26