OpenAI halves GPT-6 pricing versus the GPT-5.6 family, and Anthropic cuts Claude Opus 5.5 rates by 20 percent.
Anthropic released Claude Opus 5.5, and OpenAI released GPT-6 Sol and GPT-6 Luna about an hour later. The launches landed a day after Grok 4.7 and MiMo v2.6 Flash and Pro, packing one week with four major model releases.
OpenAI halved pricing across its new GPT-6 family relative to GPT-5.6, and Anthropic cut Claude Opus 5.5 by 20 percent to $4 per million input tokens and $20 per million output tokens. Cache read prices fell 60 percent, and GPT-6 Luna ranks among the cheapest models OpenAI has released.
Cheaper output tokens matter most for builders running agents, code review, and long document pipelines, where output volume drives cost. GPT-6 Luna at $0.10 input and $0.50 output undercuts most rivals while keeping strong application performance, so operators can revisit model routing and reprice features that were marginal at older rates.
Anthropic says Opus 5.5 targets clearer communication and stronger Blender work, so benchmark results should show whether the style complaints are resolved. GPT-5.6 carries a scheduled 25 percent price increase in November, which makes GPT-6 the cheaper option even before that hike takes effect.
What matters
- Anthropic shipped Claude Opus 5.5 and OpenAI followed within an hour with GPT-6 Sol and GPT-6 Luna.
- GPT-6 Luna costs $0.10 per million input tokens, making cheap high-volume agent workloads far more viable.
- Watch whether GPT-5.6’s scheduled November price increase holds, since GPT-6 undercuts promotional rates.
Why it matters
Watch whether GPT-5.6’s scheduled November price increase holds, since GPT-6 undercuts promotional rates.
This GenAI News article was prepared in original wording using reporting and materials published by Simon Willison’s Weblog. Source reference: https://simonwillison.net/2026/Sep/22/opus-and-sol-and-luna/.
Drafted by the GenAI News review pipeline.