Reports revealed Anthropic is internally testing Claude Opus 5.5 featuring lower token pricing

According to cnBeta, Anthropic is internally testing Claude Opus 5.5 with a greatly reduced pricing scheme of $4 per million input tokens and $20 per million output tokens, in a bid to compete with GPT-6, while Claude Opus 5.2 and Claude Fable 5.2 are also in play. The report notes Opus 5.5 touts strong performance, a major cache-usage improvement, and is planned to launch in the near term, with other players like OpenAI and Google racing their own advances (GPT-6 Sol and Gemini 4 Pro). Claude Opus 5.5 is described as part of a broader push to make top-tier AI models more affordable for enterprise use, with July 24 marked as a release milestone for Claude Opus 5 and related checkpoints mentioned.