Luna Cache Hits Broken on Codex: Users Say It's Now Pricier Than DeepSeek
Mission-Zucchini-966 · reddit · 2026-08-21
Reddit users report that cache hit rates for the Luna model on Codex team plans have gone badly wrong lately: nearly every other prompt gets a cache miss, and it's not TTL expiry—even one-minute gaps and the simplest single-line /loop prompts trigger misses.
The author notes that when DeepSeek raised prices, everyone expected Luna to be the cheaper replacement, but with the caching issues Luna now effectively costs more. They'd rather see a slight price hike with the caching fixed than a facade of "competitive pricing."
More from Models
- Ox Alpha: Free Chinese model with 1M context but high latency — bindureddy · 2026-08-21
- Legal AI startup Harvey launches first model post-trained for law on Kimi K3 — Polymarket · 2026-08-21
- Ox Alpha appears to be GLM 5.3 Flash — LinusEkenstam · 2026-08-21
- Insider confirms stealth model Ox Alpha is spectacular, resets AGI timelines — omarsar0 · 2026-08-21
- HF Report: Qwen GGUF Downloads Hit 39.6M, Dwarfing Llama's 7.5M — close_Meal6005 · 2026-08-21
- Skeptic: "Claude designed protein binders alone" was mostly just Modal credentials — charles_irl · 2026-08-21