Users allege Astra's Codex quota accounting consumes 4-5x Sol's allowance, not 2.5x
Medical-Yam3367 · reddit · 2026-09-06
Users across Chinese-language communities (Linux.do, NodeSeek) report a possible discrepancy in Astra's Codex subscription quota accounting — not just fast burn, but consumption exceeding what its published pricing implies.
- Published rates: Astra is priced at 2.5x Sol for input, cached input, and output tokens — that premium is expected.
- The anomaly: one user reports roughly 1.8x overage relative to API-priced usage; another finds the same five-hour window equals $18 of Sol usage but only $10 of Astra usage (valued at each model's API rates), consistent across four accounts.
- Aggregate: if accurate, the effective difference is 2.5 × 1.8 = 4.5x rather than 2.5x. Other data points: a Linux.do thread estimates 4-5x Sol's quota with cache hit rates of 97-98% for both models; a six-account third-party tracking comparison finds an extra 1.4-1.6x; a NodeSeek thread suggests a 1.6x accounting correction.
More from Models
- Diffusion LLM bias rankings shift with sampling configs, persist after debiasing — Bollegala · 2026-10-23
- Running ~100 DeepSeek v4.1 flash agents in parallel for under $20/month — zack_overflow · 2026-10-23
- Researcher laments open-source audio models are surprisingly subpar, even representation ones — gowthami_s · 2026-10-23
- User claims they generated video with "Claude Opus 5.5", unannounced model name surfaces — jwt0625 · 2026-10-23
- Security researchers push back as OpenAI restricts Astra access; BlackHat 2027 mocked as 'cyber model' bazaar — cyb3rops · 2026-10-23
- Mistral Large 4 Leaked on Hugging Face: 1T Params, 49B Active, Open Weights Oct 31 — mervenoyann · 2026-10-23