OpenAI and Anthropic lowered prices for several models between July and September 2026. OpenAI cut GPT-5.6 Luna’s API prices by 80% and later introduced GPT-6 Sol and GPT-6.1 Sol at lower rates, while Anthropic priced Opus 5.5 below Opus 5. The announcements point to improved inference and serving efficiency; competition from open-weight and rival models is an analysis rather than a claim established by the company announcements.

Why it matters

Lower input, output, and cached-input rates can reduce the cost of agentic coding workflows, especially for applications that reuse large prompts. Lower prices also make it easier for developers to compare or switch among competing models, increasing pressure across providers.

Key points

  • GPT-5.6 Luna’s API price fell 80% on July 30, 2026, to $0.20 per million input tokens, $0.02 per million cached input tokens, and $1.20 per million output tokens.
  • GPT-6.1 Sol’s cached-input price is $0.10 per million tokens, half the GPT-6 Sol rate and 95% below its uncached input price; OpenAI describes the model as 50% cheaper than GPT-5.6 Sol promotional pri
  • Anthropic says Opus 5.5 costs 40% less than Opus 5 on typical workloads; its input and output rates are 20% lower, while cache reads are 60% lower.

Source: The original announcement