OpenAI and Anthropic slash prices in aggressive push against open source AI models

The AI pricing war just entered a new phase. OpenAI and Anthropic have been systematically slashing prices on their mid-tier models throughout 2026, with reductions steep enough to make developers reconsider whether open source alternatives are still worth the hassle of self-hosting.

The numbers behind the price war

On July 30, OpenAI cut prices on GPT-5.6 Luna by 80%, bringing costs down to $0.20 per million input tokens and $1.20 per million output tokens. The company’s higher-end GPT-5.6 Terra also got a haircut, dropping 20% to $2 input and $12 output per million tokens.

Anthropic followed on August 10, making the introductory pricing for Claude Sonnet 5 permanent at $2/$10 per million tokens.

Then September 22 brought a second wave. OpenAI launched GPT-6 Sol and Luna at $2/$10 and $0.10/$0.50 respectively, roughly 50% lower than previous generation pricing. Anthropic countered with Claude Opus 5.5 at $4/$20 list pricing, though built-in efficiency improvements deliver around 40% effective savings for users.

Why open source is feeling the squeeze

Enterprises are increasingly splitting their workloads into tiers. Cheap proprietary models handle routine tasks like summarization, classification, and basic customer interactions. Premium models get reserved for specialized workloads where performance differences actually matter.