OpenAI unveils GPT-5.6 family: Sol, Terra, Luna with 54% better token efficiency
Sol beats Anthropic’s Fable 5 by 2.8 points while costing one-third less.
OpenAI has unveiled its GPT-5.6 family, comprising three models designed to cover a range of use cases and budgets. The flagship model, Sol, is positioned as a workhorse for complex enterprise tasks, coding, and scientific research. CEO Sam Altman claims Sol is 54% more token efficient than previous versions, particularly in AI coding tasks. The company also highlights GPT-5.6 as its strongest cybersecurity model yet, supporting defensive activities like threat modeling, code review, and blue teaming. The launch comes amid heightened scrutiny from the Trump administration over potential misuse, but OpenAI emphasizes its defensive focus.
Competitively, OpenAI is directly targeting Anthropic’s recent releases. Using the Artificial Analysis Coding Agent Index, OpenAI claims Sol scores 80—2.8 points above Anthropic’s Fable 5—while using less than half the output tokens, taking less than half the time, and costing about one-third less. Terra performs just above Fable 5, and Luna outperforms Opus 4.8. The models are available via ChatGPT, Codex, and the OpenAI API. Additionally, OpenAI introduced ChatGPT Work, a workplace companion for enterprise teams handling clerical tasks. Pricing is set at $5 input/$30 output per million tokens for Sol; Terra at $2.50/$15; and Luna at $1/$6.
- GPT-5.6 family includes Sol (flagship), Terra (mid-range), and Luna (budget) with pricing from $1 to $5 per M input tokens.
- Sol scores 80 on the Coding Agent Index, 2.8 points above Anthropic's Fable 5, using <50% tokens and costing ~33% less.
- OpenAI also launches ChatGPT Work, a desktop/web/mobile tool for enterprise clerical tasks like document drafting and spreadsheets.
Why It Matters
OpenAI’s three-tier pricing and superior coding benchmarks could reshape enterprise AI adoption and intensify rivalry with Anthropic.