Tuesday’s model news was a price move, not another frontier jump. Anthropic released Claude Opus 5.5 in the afternoon. Roughly ninety minutes later OpenAI introduced GPT-6 Sol and GPT-6 Luna.

Anthropic said Opus 5.5 matches Claude Fable 5.1 on most tasks and costs about 40 percent less than Opus 5 on typical workloads. Token pricing landed at $4 per million input and $20 per million output. On coding and knowledge-work benchmarks the new model scored at or above Fable 5.1 in the company’s published numbers.

OpenAI positioned Sol and Luna below flagship GPT-6 Astra. Sol’s API rate is $2/$10 — exactly half Opus 5.5’s listed token price. Luna sits at $0.10/$0.50. Both are described as cut from the same architecture family as Astra, optimized for cost rather than peak capability.

Key takeaways

What the price cut changes for builders

Enterprise buyers have been pressing for lower token costs as agent workloads scale. The two releases answer that pressure without a new capability race. Whether the cheaper models cannibalize higher-tier revenue or expand the addressable market is the question the next quarter’s usage data will answer.

Coverage of the launches appeared across multiple outlets within hours of the announcements. Primary sources remain the companies’ own release posts and pricing pages. For context on the broader content-stack shift, see How AI is Reshaping the Content Creator Stack in 2026.