Details
- OpenAI announces the GPT-5.6 family of models — Sol, Terra, and Luna — now rolling out in ChatGPT, Codex, and the API.
- GPT-5.6 Sol is framed as the flagship model, targeting coding, knowledge work, cybersecurity, and scientific workflows with higher intelligence and efficiency at lower token usage and cost versus prior generations.
- On Agents' Last Exam, GPT-5.6 Sol scores 53.6, beating Claude Fable 5 (adaptive) by 13.1 points and maintaining a double‑digit lead even at medium reasoning, where it also runs at roughly one-quarter the estimated cost.
- On the Artificial Analysis Coding Agent Index, Sol sets a new state-of-the-art score of 80.0, 2.8 points above Claude Fable 5, while using less than half the output tokens, taking under half the time, and costing about one-third less.
- The release introduces Ultra mode for Sol, a high‑performance setting that coordinates multiple agents in parallel to tackle demanding tasks, trading higher token use for stronger and faster results.
- GPT-5.6 significantly improves design judgment by using stronger computer‑use capabilities to inspect rendered outputs (not just code or content), identify visual and functional issues, and apply finishing touches before returning results.
- The new models enhance artifact creation across presentations, documents, and spreadsheets, working better with existing templates and allowing export into standard productivity tools for integration into real enterprise workflows.
- Availability begins immediately across ChatGPT, Codex, and the OpenAI API, with a global rollout over roughly 24 hours; in ChatGPT, Plus, Pro, Business, and Enterprise users gain access to GPT-5.6 Sol as the top-tier option for advanced work.
- Terra and Luna extend the family to more balanced and cost-efficient tiers, giving developers and enterprises multiple performance–price trade-offs for coding, agentic workflows, and productivity use cases.
Impact
This launch pushes OpenAI back to the front of the benchmark race, explicitly targeting Anthropic’s Claude Fable line on agentic coding and reasoning performance while undercutting it on effective cost per result. By pairing a premium Sol Ultra mode with cheaper Terra and Luna tiers, OpenAI broadens its stack from high‑end agents to cost‑sensitive production workloads, reinforcing its position as a default platform for enterprise AI development and intensifying pressure on rivals to ship comparably strong multi‑agent, computer‑use‑capable models.
