OpenAI has unveiled an update to its GPT-6 Sol AI model just a week after its release. The new GPT-6.1 Sol version shows results close to the leading GPT-6 Astra in complex agent tasks and programming, but costs five times less.
OpenAI has unveiled an update to its GPT-6 Sol AI model just a week after its release. The new GPT-6.1 Sol version shows results close to the leading GPT-6 Astra in complex agent tasks and programming, but costs five times less.
This is mentioned in the developer blog.
The updated model is designed for autonomous operation, writing and debugging code, analyzing complex PDF documents with tables, and executing multi-step business processes. According to the DeepSWE v1.1 test, which evaluates solving software development tasks in real code bases, GPT-6.1 Sol achieves the performance of GPT-6 Astra. At the same time, in tests for actual accuracy, the model’s error rate decreased to 4,1%.
For developers, the cost of using GPT-6.1 Sol via the API is $2 per million input tokens, $0.10 per million cached input tokens, and $10 per million output tokens. The model is already available in ChatGPT, Robota, and Codex for Plus, Pro, Business, Enterprise, and Edu tiers.
At the same time, OpenAI introduced accelerated versions of GPT-6 Astra Ultrafast and GPT-6.1 Sol Ultrafast, which run up to 8 times faster than the base Astra. They are available to subscribers of the new Pro plan, which costs $500 per month.
OpenAI has previously expanded its GPT-6 family of models with two new modifications, GPT-6 Sol and GPT-6 Luna, designed to deliver high speed and cost savings for complex programming, automation, and data analysis tasks.
