$20-tier user reports Astra Light exhausts usage in 5-6 prompts, unlike Sol
HousingLatter6637 · reddit · 2026-09-10
A Reddit user on the $20 plan says Astra Light runs out of usage before or between the 5th and 6th prompt, contradicting OpenAI's claim that Astra consumes fewer tokens than Sol — they found Sol medium/high lasted longer in practice.
More from Models
- Screenshot surfaces rare admission of 72-hour KV cache limits in V4-era architecture — zephyr_z9 · 2026-09-10
- User says OpenAI shut down his protein design project, pivoting to open-weight models — Terminator857 · 2026-09-10
- Devs call for standardized "model performance across harnesses" evals — zainhas · 2026-09-10
- Rumored DeepSeek V4.1 Flash details point to asymmetric-activation MoE with much lower cost — gaganghotra_ · 2026-09-10
- DeepSeek V4.1 Flash's reasoning_effort scales output quality and tokens ~linearly in tests — zainhas · 2026-09-10
- AI Researchers Buzz Over a Brand-New Kind of Encoder-Decoder Model — antoine_chaffin · 2026-09-10