OpenAI blinks in face-off with Chinese rivals, drops pricing for some models up to 80%

US AI company OpenAI has conceded to fierce competition from Chinese counterparts by significantly reducing prices for some of its cutting-edge GPT-5.6 models. In a bid to hold onto its market share, OpenAI has slashed the cost of its lightweight GPT-5.6 Luna model by up to 80%, bringing API fees down to $0.20 per million input tokens and $1.20 per million output tokens.

The San Francisco-based firm also cut rates for its mid-tier GPT-5.6 Terra model by 20%, now priced at $2 per million input tokens and $12 per million output tokens. Additionally, OpenAI introduced a "fast mode" for its flagship GPT-5.6 Sol model, promising up to a 2.5 times performance boost. These drastic price cuts have catapulted GPT-5.6 Luna into the top tier of intelligence-per-dollar rankings, surpassing Chinese rivals like Zhipu AI’s GLM-5.2 and MiniMax’s M3.

This move underscores the growing challenge Western technology giants face in an era where low-cost, high-performance Chinese models offer aggressive pricing and rapidly improving capabilities. Beijing-based Moonshot AI recently launched API access to its open-weight Kimi K3 model, which has outperformed OpenAI’s GPT-5.6 Sol and Anthropic’s Claude Fable 5 on certain benchmarks.

However, Moonshot AI's new pricing for Kimi K3 remains higher than OpenAI's updated GPT-5.6 Terra and Luna tiers. Despite OpenAI's reductions, rival Chinese models such as Zhipu AI’s GLM-5.2 and MiniMax’s M3 still offer competitive pricing.