OpenAI Launches GPT-6.1 Sol With Near-Astra Performance at Lower Cost
The new model targets agentic coding and professional work while cutting API costs well below GPT-6 Astra.
OpenAI has launched GPT-6.1 Sol, a new model aimed at agentic coding, computer use, and professional work. The company says the model brings performance close to GPT-6 Astra while using a much lower-cost pricing structure. The launch comes shortly after the original GPT-6 Sol and Luna release, adding another step to OpenAI's rapidly expanding GPT-6 lineup.
GPT-6.1 Sol Targets Near-Astra Performance
OpenAI describes GPT-6.1 Sol as an upgrade over GPT-6 Sol for workloads that require sustained reasoning and tool use. Its main focus is practical work such as software development, computer-use tasks, and other professional workflows that can be handled through agents.
In OpenAI's published evaluations, GPT-6.1 Sol reached the same result as GPT-6 Astra on the DeepSWE v1.1 software-engineering benchmark. The company also reports improvements over the earlier GPT-6 Sol at lower reasoning effort. These are OpenAI's own benchmark results, so they should be read as model-maker evaluations rather than independent testing.
The Cost Difference Is the Main Change
The standard API price for GPT-6.1 Sol is $2 per million input tokens and $10 per million output tokens. GPT-6 Astra is priced at $10 per million input tokens and $50 per million output tokens, putting Sol at one-fifth of Astra's standard input and output rates.
That pricing makes GPT-6.1 Sol relevant for applications that need capable coding or agentic behavior but cannot justify Astra's higher per-token cost. Developers comparing model economics can also use Saganote's AI API Pricing 2026 guide to see how OpenAI's rates compare with other major models.
Where GPT-6.1 Sol Fits in OpenAI's Lineup
GPT-6.1 Sol sits below Astra in OpenAI's model lineup while targeting some of the same types of professional workloads. OpenAI's current model documentation describes Sol as a model for complex coding and agentic workflows, with support for functions, web search, file search, and computer use.
The model is being rolled out in ChatGPT Work and Codex, with availability expanding across paid and organizational plans. Developers can also access it through the OpenAI API. The model's positioning gives teams another option between lower-cost models and the more expensive Astra tier.
For readers tracking OpenAI's recent model changes, the launch follows earlier coverage of GPT-6 Astra's developer pricing and Codex changes and the broader GPT-5.6 Sol launch. It also comes as Anthropic has expanded its own agentic coding offering with Claude Sonnet 5.
What Happens Next
OpenAI's main claim for GPT-6.1 Sol is that developers can get much of Astra's capability profile at a substantially lower API cost. Independent evaluations will be useful for determining how closely that positioning holds across coding, computer use, and broader professional workloads.
OpenAI's official announcement is available at https://openai.com/index/introducing-gpt-6-1-sol/. The launch also adds another data point to the wider OpenAI vs Anthropic vs Google DeepMind comparison, as major AI labs continue adjusting model capability and pricing for agentic workloads.