The GPT-5.6 series has officially launched, releasing three models at once: Sol, Terra, and Luna. Sol costs $5 per million input tokens and $30 per million output tokens; Terra costs $2.5 and $15; Luna costs $1 and $6. All three models are fully available starting today, rolling out across ChatGPT, Codex, and the API within 24 hours. The standalone Codex app is being discontinued, with its capabilities integrated into ChatGPT. Under pressure, Anthropic voluntarily reset Claude's usage quotas. In the Agents' Last Exam benchmark, GPT-5.6 Sol scored 53.6 points, 13.1 points higher than Claude Fable 5. OpenAI stated that even with medium-level reasoning enabled, Sol outperforms Fable 5 by 11.4 points at roughly a quarter of the cost, and Terra and Luna surpass Fable 5 at about one-sixteenth the cost. On the Artificial Intelligence Index, Sol at full reasoning mode is 1 point behind Fable 5 but completes tasks 61% faster at roughly half the cost. Billing rules have been adjusted: writing to prompt cache is charged at 1.25x the regular input rate, reading from cache costs 10% of the regular price, and cached content is retained for at least 30 minutes.