OpenAI has released GPT-6.1 Sol — a model that, according to the company, nearly matches the flagship GPT-6 Astra in capabilities while costing significantly less. We break down the differences and who the new model is best suited for.
What Is GPT-6.1 Sol
GPT-6.1 Sol is a continuation of the GPT-6 lineup, released less than a month ago alongside the flagship Astra. According to OpenAI, Sol has practically caught up with Astra in response quality across most typical tasks while offering a lower cost of use through the API.
This is the same approach OpenAI already used in the GPT-5.6 lineup — at that time, the company also released several versions of the model at different tiers (Sol, Terra, Luna) for different budgets and use cases.
How Sol Differs from Astra
The main difference is not in the capabilities themselves, but in the balance between price and performance. Astra remains the technical flagship, with record results on difficult benchmarks such as FrontierMath, and it is the model that received a “critical” cyber capability classification, which is why access to its most advanced capabilities is restricted.
Sol is aimed at the mass market — developers and businesses that need high-quality responses at a reasonable price, without requiring extreme performance on highly specialized tests that are more relevant to researchers than to everyday use.
Where GPT-6.1 Sol Is Already Available
The model is available through the OpenAI API immediately after release. This contrasts with Astra’s phased rollout, where access was initially limited to verified partners in the Daybreak cybersecurity program before gradually expanding to regular ChatGPT users.
Why This Matters for the Market
The release of Sol continues a trend that began taking shape in the summer: leading AI labs are releasing cheaper versions of their flagship models one after another in response to competition from lower-cost Chinese alternatives such as DeepSeek, Kimi and Qwen. Anthropic did the same with Claude Opus 5 and later with Fable 5.1 — both models cost significantly less than the top versions of their respective lineups while offering similar quality.
In Brief: The Key Points
- OpenAI released GPT-6.1 Sol — a cheaper version of the flagship GPT-6 Astra
- According to the company, Sol nearly matches Astra in quality on typical tasks
- Astra remains the technical flagship with record results on difficult benchmarks and restricted access due to its “critical” cyber-risk classification
- Sol is aimed at developers and businesses that need quality at a reasonable price
- The model is available through the API immediately, without the phased restricted rollout used for Astra
- The release continues the trend of lowering flagship model costs in response to competition from China
FAQ
What is GPT-6.1 Sol? A cheaper version of OpenAI’s flagship GPT-6 Astra model that, according to the company, nearly matches it in quality on most typical tasks.
How does Sol differ from Astra? Astra is the technical flagship with record results on difficult benchmarks and restricted access due to its critical cyber-risk classification. Sol is designed for widespread use at a lower price.
Where is GPT-6.1 Sol available? Through the OpenAI API immediately after release, without the phased access restrictions that were applied to Astra.
Why is OpenAI releasing cheaper versions of its models? Because of competition from lower-cost Chinese AI models — DeepSeek, Kimi and Qwen — which are becoming increasingly popular among companies seeking to control AI spending.
The article was prepared by the TechVisor team — practical IT media for people.
