source: techcrunch ai: openai launches its new family of models with gpt-5.6
level: business
openai introduced gpt-5.6, a new family of models with three tiers: sol, terra, and luna. sol is the most capable, terra is intermediate, and luna is budget-friendly. the company says these models improve efficiency and cost, with sol being 54% more token efficient for coding tasks compared to earlier versions. openai also highlights cybersecurity performance, calling it the strongest model yet for defensive tasks like threat modeling and code review.
the launch includes chatgpt work, a tool for enterprise teams to handle clerical tasks across desktop, web, and mobile. openai positions gpt-5.6 against anthropic's models, claiming sol outperforms fable 5 on the artificial analysis coding agent index by 2.8 points while using fewer tokens and costing less. terra matches fable 5, and luna beats opus 4.8. pricing per million tokens is $5/$30 for sol, $2.50/$15 for terra, and $1/$6 for luna.
the release follows similar announcements from spacexai and meta, but openai's messaging targets anthropic's enterprise focus. the models are available via chatgpt, codex, and the api. openai emphasizes token efficiency and cost savings, aiming to attract developers and businesses. the cybersecurity capabilities were previously under scrutiny by the trump administration, but openai frames them as defensive tools.
why it matters: gpt-5.6 offers cheaper, more efficient models for coding and enterprise tasks, directly competing with anthropic and potentially lowering costs for ai development.
source: techcrunch ai: openai launches its new family of models with gpt-5.6