MUMBAI: The AI price war is getting a little more economical. OpenAI is cutting the cost of using its GPT-5.6 Sol model for developers for the next three months, lowering API prices as competition intensifies from Anthropic and Chinese AI models.
Under the revised pricing, standard short-context use of GPT-5.6 Sol will cost $4 per 1 million input tokens, down from $5, while output tokens will cost $20 per 1 million, compared with $30 earlier. That represents a 20 per cent reduction in input costs and roughly 33 per cent in output costs, with the overall reduction exceeding 20 per cent.
The lower rates will apply to OpenAI’s API and will also be available across eligible credit-based plans for ChatGPT Work and Codex, its coding tool. Subscription prices for Pro, Plus and Business users, however, will remain unchanged.
The latest cut follows a broader pricing push from OpenAI. Late last month, the company reduced the price of its mid-tier GPT-5.6 Terra by 20 per cent, while its lower-cost GPT-5.6 Luna received an 80 per cent reduction.
The move comes as developers have more models to choose from and AI companies compete increasingly on both performance and cost. Anthropic’s Claude Fable 5 is priced at $10 per 1 million input tokens and $50 per 1 million output tokens, while Claude Opus 5 costs $5 per 1 million input tokens and $25 per 1 million output tokens.
For developers running AI applications at scale, the difference is not merely decimal. Token costs can quickly add up when models are processing millions or billions of inputs and outputs, making pricing an increasingly important part of model selection.
OpenAI’s three-month reduction therefore puts more emphasis on the economics of building with GPT-5.6 Sol, while signalling that the battle for developers is being fought not just with smarter models, but with cheaper ones too.
