OpenAI's GPT-6.1 Sol nears GPT-6 Astra on agent work at a fifth of the price
Serguey Asael Shinder1 min readOpenAI has released GPT-6.1 Sol, an upgrade to GPT-6 Sol that the company says nearly matches GPT-6 Astra on agentic coding, computer use and professional work at one-fifth of Astra's standard input and output token prices. Cached input costs $0.10 per million tokens, which OpenAI puts at 95% below standard input pricing and half of GPT-6 Sol's cached rate.
On DeepSWE v1.1, which tests long-horizon software engineering in real codebases, OpenAI reports that GPT-6.1 Sol matches Astra at roughly one-fifth of the cost and beats GPT-6 Sol's best score by 6.4 percentage points at a lower reasoning effort. On OSWorld 2.0's offline set for computer use it gains seven points over GPT-6 Sol at maximum effort and comes within 2.1 points of Astra at about one-seventh of the cost per task. On Terminal-Bench Science it more than doubles GPT-6 Sol's score; Astra still scores highest there at 68.1%, and OpenAI recommends it for the hardest research tasks.
The announcement also compares against competitors. On GDP.pdf, a benchmark of questions about complex professional PDFs, and on AutomationBench, a test of multi-step business workflows across 47 tools, OpenAI reports GPT-6.1 Sol ahead of Anthropic's Opus 5.5 at lower cost. On factuality, the share of answers containing an error on deliberately difficult prompts fell from 11.4% to 7.7% at low reasoning effort, a set OpenAI says is not representative of typical usage.
Why it matters
For teams running agents, the headline is the cached-input price: agents re-send long contexts on every step, so a cheaper cache changes the cost of a whole workflow more than a cheaper list price does. The benchmark numbers are OpenAI's own and its competitor comparisons use its chosen settings; independent runs will show how much of the gap holds.