OpenAI and Anthropic shipped on Sep 22. OpenAI launched GPT-6 Sol and Luna at half the price of the GPT-5.6 promo and added a dashboard and longer reuse for GPT-6 prompt caching. Anthropic shipped Claude Opus 5.5 with cheaper runs.
In the news
GPT-6 Sol and Luna launch at half price
OpenAI launched GPT-6 Sol and Luna on Sep 22. Sol is priced at $2 per million input and $10 per million output. Luna is $0.10 per million input and $0.50 per million output. Both are half the price of the GPT-5.6 promo. AutomationBench puts Sol at 33.2 percent at $0.27 per task. DeepSWE v1.1 at max effort hits 68.8 percent for Sol.1
Claude Opus 5.5 ships with cheaper runs
Anthropic shipped Claude Opus 5.5 on Sep 22. It is the first model in the 5.5 line. Pricing is $4 per million input and $20 per million output. Cache reads cost $0.20 per million. Anthropic says it runs 40 percent cheaper than Opus 5 on typical workloads. Output is more than 30 percent faster than Opus 5. Terminal-Bench 4.0 scores 66.4 percent.1
GPT-6 prompt caching gets a dashboard and breakpoints
OpenAI improved prompt caching for GPT-6 on Sep 22. The reuse window is now 30 minutes. A Prompt Caching Dashboard and diagnostics tool are available. Explicit cache breakpoints and prewarming are added. Reasoning effort can change without breaking the cache.1
