[SAN FRANCISCO, United States] — OpenAI has announced a reduction in the API pricing of its GPT-5.6 Sol model, with developers set to pay less for standard short-context usage over the next three months. The reported price reduction applies to API access and eligible credits connected with ChatGPT Work and Codex, while ChatGPT Pro, Plus and Business subscription prices remain unchanged.
- Key Highlights:
- GPT-5.6 Sol API input pricing is reported at $4 per 1 million tokens, down from $5.
- Output pricing has fallen to $20 per 1 million tokens from $30.
- The reduced pricing is set to apply for the next three months.
- OpenAI’s subscription prices for Pro, Plus and Business users remain unchanged.
GPT-5.6 Sol API prices reduced
The pricing change targets developers and businesses that access OpenAI’s frontier model through its application programming interface. Under the revised rates cited in the supplied material, standard short-context usage of GPT-5.6 Sol costs $4 per 1 million input tokens and $20 per 1 million output tokens.
The earlier rates were listed at $5 per 1 million input tokens and $30 per 1 million output tokens. That represents a reduction in both input and output costs, with the larger reduction applying to generated output.
At the exchange rates used in the supplied report, the new input price is approximately ₹380 per 1 million tokens, while output costs around ₹1,900 per 1 million tokens. The previous rates were approximately ₹480 and ₹2,870 respectively.
Who gets the lower GPT-5.6 Sol pricing?
The reduction is focused on API usage rather than consumer ChatGPT subscriptions. OpenAI said the lower rates are rolling out across eligible plans for credits associated with its agentic AI product ChatGPT Work and its coding tool Codex.
For developers, API pricing is particularly significant because token usage can become a major operating cost when AI models are integrated into applications, automated workflows and coding systems. Lower per-token rates can therefore affect the cost of running larger volumes of model-powered tasks.
However, the supplied material makes clear that the change does not extend to the listed ChatGPT subscription plans. Users on Pro, Plus and Business subscriptions will continue to pay the existing subscription prices.
OpenAI expands model pricing cuts
The latest move follows other pricing changes across the GPT-5.6 model family. Late last month, OpenAI reportedly reduced the price of its smaller models and cut the price of its mid-tier GPT-5.6 Terra model by 20%.
The lower-cost GPT-5.6 Luna model also received a much larger price reduction, with its price reportedly falling by 80%. Together, the changes indicate a broader effort to reduce the cost of accessing different levels of OpenAI’s model lineup.
The pricing moves come as OpenAI faces increasing competition from other AI companies and models. The supplied material specifically points to competition from Anthropic and Chinese AI models as part of the wider market environment surrounding the changes.
How GPT-5.6 Sol compares with Anthropic models
The supplied pricing comparison places OpenAI’s revised GPT-5.6 Sol rates below the listed prices of Anthropic’s frontier models. Anthropic’s Claude Fable 5 is listed at $10 per 1 million input tokens and $50 per 1 million output tokens.
Anthropic’s Claude Opus 5 is listed at $5 per 1 million input tokens and $25 per 1 million output tokens. On the figures provided, the revised GPT-5.6 Sol rates are lower for both input and output usage than those listed alternatives.
For developers choosing an AI model, token pricing is only one part of the decision. The actual cost of an application can also depend on how much input and output a workflow generates, making the difference between input and output rates important for high-volume deployments.
ChatGPT for Teens also launched
OpenAI has also introduced a teenager-focused version of ChatGPT designed to support learning and help younger users use AI with additional protections. The company said the experience is intended to help teenagers learn, deepen their understanding and use AI with confidence.
According to OpenAI, the system is designed to activate the teen-focused experience when it estimates that a new user is between 13 and 17 years old. The company has also highlighted parental controls, its under-18 model guidance, the Teen Safety Blueprint and age-prediction work as part of its broader approach to teen safety.
OpenAI said its approach is guided by four commitments: putting teen safety first, encouraging real-world support, treating teenagers according to their age and being transparent about how its systems should behave.
Pricing and product strategy move together
The reported API reduction comes alongside changes across OpenAI’s model portfolio and new product initiatives. The company has continued to differentiate its models by capability, cost and intended use, while expanding access through its API, ChatGPT and coding products.
For developers, the three-month pricing window described in the supplied material makes the revised GPT-5.6 Sol rates particularly relevant for teams assessing the economics of AI-powered applications and automated workloads.
Praveen Yadav is the Founder and Content Creator of The Nation Bulletin, an independent digital news platform focused on delivering timely, reliable and meaningful news from India and around the world.
