OpenAI launches its new family of models with GPT-5.6
Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated.
OpenAI's new GPT-5.6 Sol model sets a new coding benchmark state-of-the-art at 80 on the Coding Agent Index, outperforming Anthropic's Fable 5 while using less than half the output tokens, running twice as fast, and costing one-third less. For production systems, this dramatically lowers the latency and cost barriers of running autonomous coding agents, while the budget-friendly Luna variant at $1 input / $6 output per million tokens makes high-volume workflow automation significantly more viable. You can immediately deploy these models via the OpenAI API across three tiers—Sol, Terra, and Luna—to optimize your pipelines for speed, cost, or reasoning depth.
GPT-5.6's Sol variant achieves 54% better token efficiency for coding tasks while outperforming Anthropic's Fable by 2.8 points on the Coding Agent Index, at one-third the cost. This lets teams deploy higher-quality AI coding assistance with lower latency and budget impact, directly shifting cost-benefit calculations for production pipelines. The Terra and Luna options further allow tiered performance/cost tradeoffs, making it easier to right-size models for specific workloads.