OpenAI introduces GPT-6.1 Sol: an upgrade to GPT-6 Sol, near-Astra on stated tasks, standard API $2 / $10
0 reads
- Published
- 2026-09-30
- Type
- Model release
- Organization
- OpenAI
- Source
- OpenAI
Why it matters
【What it is】
OpenAI published “Introducing GPT-6.1 Sol” (https://openai.com/index/introducing-gpt-6-1-sol/). It upgrades GPT-6 Sol and nearly matches GPT-6 Astra on agentic coding, computer use, and professional work, at one-fifth of Astra’s standard input and output prices.
API id gpt-6.1-sol. From today: Plus, Pro, Business, Enterprise, and Edu in ChatGPT Work and Codex; not in Chat yet. September 29, 2026. This is not the September 22 GPT-6 Sol and Luna release.
【Verifiable】
Per million tokens: input $2, cached input $0.10, output $10. Cached input is 95% below this model’s input price and 50% below GPT-6 Sol’s cached input. Astra’s standard price is $10/$50 (https://openai.com/index/gpt-6-astra/). Opus 5.5 is $4/$20 (https://www.anthropic.com/claude-opus-5-5/). $2/$10 matches GPT-6 Sol and is one-fifth of Astra.
DeepSWE v1.1 matches Astra at about one-fifth of the cost and beats GPT-6 Sol’s best by 6.4 percentage points at lower effort and cost.
Terminal-Bench Science 0.1, maximum effort, averages $5.47 per task, versus $23.21 for Opus 5.5 and $23.80 for Astra (over 75% lower). It more than doubles GPT-6 Sol at less than half the cost. Astra’s cited high is still 68.1%. The page says the hardest scientific research should stay on Astra.
GDP.pdf is higher than Opus 5.5 with fallbacks at less than half the cost per task, and near Astra at about one-fifth the cost. AutomationBench 1.0.6 is 2.2 percentage points above Opus 5.5 at medium effort, at about a third of the cost, and 4.8 percentage points above GPT-6 Sol at the same setting. Broken-search non-disclosure at maximum effort is 2.1% versus 4.9% Sol, 1.5% Astra, 28.7% Luna.
【Our read】
$2/$10 matches GPT-6 Sol, is half of Opus 5.5 ($4/$20) on input and output, and one-fifth of Astra ($10/$50). That is list price, not a per-task bill.
“Near-Astra” is agentic coding, computer use, and professional work, not the hardest science. Versus September 22 Sol, $2/$10 is unchanged and cached input is 50% lower.
Related developments
Aggregated by AGI Pulse. Titles and links only; no full text is republished. Editorial policy