OpenAI launches GPT-6.1 Sol with near-Astra performance at one-fifth the cost: What you get

OpenAI has launched GPT-6.1 Sol, an upgraded version of GPT-6 Sol that is designed to deliver performance closer to the company’s more powerful GPT-6 Astra model without the same cost. The new model targets tasks beyond simple chatbot queries, including coding, professional workflows, computer-based tasks and scientific research.

OpenAI says GPT-6.1 Sol can approach Astra’s performance across several of these areas at around one-fifth of the cost. This could make it particularly useful for developers building AI agents that need to handle complex, multi-step tasks. The model also brings improvements in factual accuracy, transparency and safety, while its pricing could give developers more room to run AI workloads at scale.

Must Read: OpenAI introduces Dots: Always-on AI agents that work 24/7 and act on your behalf

GPT-6.1 Sol targets complex work at a lower cost

For coding, GPT-6.1 Sol matches GPT-6 Astra on DeepSWE v1.1 at roughly one-fifth of the cost and beats GPT-6 Sol’s score by 6.4 percentage points. On GDP.pdf, which tests performance on complex document-based tasks, it scored higher than Opus 5.5 with fallbacks at less than half the cost per task, while approaching Astra’s performance at about one-fifth of the cost.

The model also scored 2.2 percentage points higher than Opus 5.5 on AutomationBench at medium reasoning effort, at roughly one-third of the cost. Its score was 4.8 percentage points higher than GPT-6 Sol at the same reasoning level.

Must Read: OpenAI cancels new AI model launch last-minute over safety concerns: All details

For computer use, GPT-6.1 Sol improved by seven percentage points over GPT-6 Sol on OSWorld 2.0 and came within 2.1 percentage points of Astra at roughly one-seventh of the cost. In scientific research, it more than doubled GPT-6 Sol’s Terminal-Bench Science score. At maximum reasoning effort, the average cost was $5.47 per task, compared with $23.21 for Opus 5.5 and $23.80 for Astra. Astra, however, recorded the highest score at 68.1%.

What GPT-6.1 Sol costs and where you can use it

The model reduces factual errors from 11.4% to 7.7% at low reasoning effort. OpenAI notes that this test used deliberately difficult prompts and does not represent typical usage. It also showed improvements in transparency, handling user intent and following safety restrictions, with no attempts to bypass automated safety review.

GPT-6.1 Sol is available to Plus, Pro, Business, Enterprise and Edu users in ChatGPT Work and Codex, but not yet in Chat. Developers can access it through the API at $2 per million input tokens, $0.10 for cached input and $10 for output tokens. The cached-input price is 95% lower than standard input pricing and 50% lower than GPT-6 Sol’s cached-input price. An Ultrafast version is also coming to Codex, offering up to eight-times faster token generation than the standard speed