OpenAI has introduced GPT-6.1 Sol, a new model designed to bring performance close to its flagship GPT-6 Astra while offering substantially lower API pricing.
According to OpenAI, GPT-6.1 Sol is built for complex coding, computer use and professional workflows. The company describes it as a model that balances intelligence, speed and cost for users handling demanding AI tasks.
The biggest change is its pricing. GPT-6.1 Sol costs $2 per million input tokens and $10 per million output tokens under standard API pricing. GPT-6 Astra is priced at $10 per million input tokens and $50 per million output tokens, making Sol one-fifth the standard per-token cost of Astra.
This pricing structure could make advanced AI capabilities more accessible for developers and businesses that run large numbers of model requests. Lower token costs can be particularly relevant for applications involving repeated coding tasks, long workflows and automated computer interactions.
OpenAI says GPT-6.1 Sol delivers near-Astra performance for complex work rather than positioning it as an identical replacement for Astra. Its official model guidance recommends Sol when users want a combination of strong reasoning capabilities and lower costs.
The model also supports a broad range of tools through OpenAI’s Responses API. These include web search, file search, image generation, code interpreter, hosted shell, computer use, MCP and other capabilities designed for agentic workflows.
GPT-6.1 Sol has a context window of up to 1.05 million tokens and can generate up to 128,000 output tokens. It supports several reasoning settings, ranging from low through maximum effort, giving developers more control over how much reasoning is applied to a task.
OpenAI’s latest safety evaluation also shows that GPT-6.1 Sol has improved substantially over the earlier GPT-6 Sol model in several areas. On HealthBench Professional, for example, GPT-6.1 Sol recorded a length-adjusted score of 64.2 compared with 60.8 for GPT-6 Sol, while GPT-6 Astra scored 64.7.
The company also reports that GPT-6.1 Sol performed within 0.5 percentage points of GPT-6 Astra across four HealthBench evaluations, although performance varies depending on the specific task and benchmark. In cybersecurity-related evaluations, the newer Sol model still remained below Astra on some measures.
For developers already using GPT-6 Sol, OpenAI now recommends reviewing its migration guidance before moving applications to GPT-6.1 Sol. The newer model is presented as the latest version of the Sol family and is specifically designed for complex coding and agentic workloads.
OpenAI’s broader GPT-6 model lineup now includes Astra, Sol and Luna, with each model targeting a different balance between intelligence, speed and cost. Astra is positioned for the most demanding work, Sol for a balance of capability and affordability, and Luna for faster, cost-sensitive workloads.
For companies building AI agents, software development tools and professional automation systems, GPT-6.1 Sol’s lower pricing could be particularly significant. The model combines advanced tool support with a large context window while reducing the per-token expense compared with GPT-6 Astra.
