GPT-6.1 Sol: near top performance at a fifth of the price

OpenAI's DevDay update targets coding, computer use and professional tasks.

ChatGPTModelsOpenAI
GPT-6.1 Sol

At DevDay 2026 on 29 September, OpenAI presented GPT-6.1 Sol, an update built for coding, computer use and professional tasks. OpenAI says it comes close to its top model, GPT-6 Astra, on several evaluations while costing one fifth as much.

Key facts

  • Announced at OpenAI DevDay 2026 (29 September 2026).
  • An update to GPT-6 Sol, aimed at coding, computer use and professional work.
  • OpenAI says it approaches GPT-6 Astra on several evaluations at one fifth of Astra’s standard token prices.
  • Cached input costs $0.10 per million tokens.
  • Available in the API, ChatGPT Work and Codex, for Plus, Pro, Business, Enterprise and Edu.

What happened

One of the model announcements at DevDay 2026 was GPT-6.1 Sol. It updates GPT-6 Sol with a focus on three areas: writing and fixing code, operating software through computer use, and professional tasks.

OpenAI compared it with GPT-6 Astra, its highest performing model. According to OpenAI, GPT-6.1 Sol comes close to Astra on several evaluations while its standard input and output prices are one fifth of Astra’s. Cached input, used when the same context is sent again, costs $0.10 per million tokens.

GPT-6.1 Sol was not the only news of the day. According to InfoQ, OpenAI also announced computer use for the Agents API, cloud-based Codex environments, a Decisions API and new plugin capabilities for ChatGPT.

The model is available in three places: the API for developers, ChatGPT Work and Codex, OpenAI’s coding agent. In ChatGPT it comes with the Plus, Pro, Business, Enterprise and Edu plans.

Why it matters

Price is what decides which model ends up inside everyday tools. When a model with near top results costs a fifth as much, companies can use it for more tasks and more people, and it tends to reach regular plans and apps faster.

For users, the quality of the result still depends heavily on the prompt. Professional tasks in particular reward a clear structure: a role, the context, the exact output and what to do when information is missing. A strong model with a vague prompt gives confident but generic answers. The same model with a precise prompt gives work you can actually use.

How to try it

Job hunting is one of the most searched uses of AI right now, and it is a good test for professional prompts. Open our free prompt Analyze a job ad and paste a job advert you are interested in. In the brackets, add a few lines about your background, for example your current role and years of experience.

The prompt breaks the ad into real requirements, hidden signals and questions to ask, so you can decide whether and how to apply. Run it in ChatGPT, Claude or Gemini.

Check the answer for three things: the requirements match what the ad really says, the “hidden signals” are presented as interpretations rather than facts, and the questions are specific to that role. If the model adds requirements that are not in the ad, ask it to quote the line it is based on.

Frequently asked questions

What is GPT-6.1 Sol?

It is an update to GPT-6 Sol announced at OpenAI DevDay 2026, designed for coding, computer use and professional tasks. OpenAI says it approaches GPT-6 Astra on several evaluations.

How much does GPT-6.1 Sol cost?

OpenAI says its standard token prices are one fifth of GPT-6 Astra's, and cached input costs $0.10 per million tokens.

Where can I use GPT-6.1 Sol?

It is available through the OpenAI API, in ChatGPT Work and in Codex.

More news

All news

Today's trends

Every day we update what people search for, country by country.

Open the trends