OpenAI Launches GPT-5.6 With Three-Tier Lineup
OpenAI launched GPT-5.6, a new model family with three variants: Sol (workhorse), Terra (intermediate), and Luna (budget). The release is available across ChatGPT, Codex, and the OpenAI API.
Pricing
Sol is priced at $5 input / $30 output per million tokens. Terra comes in at $2.50 / $15. Luna is the cheapest at $1 / $6.
The Benchmark Claims
CEO Sam Altman says Sol is 54% more token efficient than previous versions for AI coding tasks. On the Artificial Analysis Coding Agent Index, Sol scores 80, which OpenAI claims is 2.8 points above Anthropic's Fable 5.
OpenAI also claims Sol uses less than half the output tokens, takes less than half the time, and costs about one-third less than Fable 5. Terra is said to perform just above Fable 5. Luna reportedly outperforms Opus 4.8.
These are OpenAI's numbers. Independent verification is pending.
Cybersecurity Positioning
OpenAI is calling GPT-5.6 its strongest cybersecurity model yet. Supported use cases include threat modeling, code review and patching, and blue teaming. The pitch is frontier-level security performance at lower token counts.
ChatGPT Work
OpenAI also shipped ChatGPT Work, a workplace companion tool for enterprise teams. It runs on desktop, web, and mobile.
Context
The head-to-head framing against Fable 5 is specific enough to be falsifiable. A 2.8-point coding lead, half the tokens, half the time, one-third the cost. If those numbers hold under independent testing, it's a real efficiency gap. If they don't, the benchmark marketing team has been busy.
The three-tier structure also mirrors how Anthropic and Google have organized their recent releases. The naming pattern is different but the strategy is the same: price-discriminate across use cases without fragmenting the brand.
Source: Techcrunch