OpenAI turns GPT‑5.6 into a three-tier model portfolio
OpenAI released Sol, Terra and Luna as a three-tier GPT-5.6 portfolio, then cut Luna API prices by 80% and Terra prices by 20% on 30 July. Fast mode replaced Priority Processing for Sol, promising up to 2.5 times standard speed at twice the price while preserving priority-tag compatibility.
Why it matters: The routing decision now spans quality, latency and sharply different unit economics rather than one model ranking. Existing priority traffic also changed processing semantics without requiring a request rewrite, so cost and latency baselines need fresh…