OpenAI Launches GPT-6.1 Sol Model With Reduced API Costs

01 What happened

OpenAI has released GPT-6.1 Sol, an upgraded model designed to match the capabilities of GPT-6 Astra in agentic coding and professional workflows while lowering token costs.

02 Key details

• OpenAI claims the model matches GPT-6 Astra's performance on DeepSWE v1.1 coding benchmarks at one-fifth of the cost.
• Cached input tokens for the new model are priced at $0.10 per million, representing a 95% reduction from standard input pricing.
• The model demonstrates a 32% reduction in factual errors compared to GPT-6 Sol during tasks requiring low reasoning effort.

03 Why it matters

The release provides a more economically viable option for developers building complex AI agents and software engineering tools that require high reasoning capabilities.

04 Who it matters to

Software developers, high-performance computing specialists and enterprise AI customers.