OpenAI is introducing GPT-6.1 Sol, an upgrade to GPT-6 Sol that brings it close to GPT-6 Astra‘s performance on agentic coding, computer use, and professional work at one-fifth of Astra’s standard input and output token prices. Cached input pricing drops to $0.10 per million tokens — 95% below standard input pricing and 50% below GPT-6 Sol’s cached input pricing — which the company frames as room for developers to build and run capable agents that reuse context across requests.

The release reports benchmark results across five areas. On DeepSWE v1.1, which evaluates complex software-engineering tasks in real codebases, GPT-6.1 Sol matches GPT-6 Astra at roughly one-fifth the cost and eclipses GPT-6 Sol’s best score by 6.4 percentage points at lower reasoning effort and cost. On GDP.pdf, which measures how accurately models answer professional questions from complex PDFs (tables, charts, diagrams, fine print) drawn from finance, healthcare, legal, and seven other professional domains, it scores higher than Opus 5.5 with fallbacks at less than half the cost per task, and approaches GPT-6 Astra‘s state-of-the-art performance at roughly one-fifth the cost per task. On AutomationBench 1.0.6, where agents run end-to-end workflows across 47 tools in sales, marketing, operations, support, finance, and HR, GPT-6.1 Sol scores 2.2 percentage points above Opus 5.5 at medium reasoning effort at roughly a third of the cost, and 4.8 points above GPT-6 Sol at the same setting. The source cautions that the Claude Fable 5.1 datapoint understates its actual cost because it omits fallbacks, which occurred on roughly 40% of tasks.

On OSWorld 2.0’s offline set for demanding computer-use workflows, GPT-6.1 Sol outperforms GPT-6 Sol by seven percentage points at maximum reasoning effort at less than half the cost, landing within 2.1 points of Astra’s score at roughly one-seventh the cost per task. On Terminal-Bench Science 0.1, covering data analysis, simulation, and theorem proving, it more than doubles GPT-6 Sol’s score at maximum reasoning effort at less than half the cost per task — $5.47 per task on average versus $23.21 for Opus 5.5 and $23.80 for Astra, more than 75% lower than either. The source is explicit that GPT-6 Astra still achieves the highest scientific score among tested models at 68.1% and should be used for the most difficult scientific research tasks.

Factuality also improves, most at low reasoning effort: the share of responses containing at least one factual error drops from 11.4% to 7.7%, a reduction of roughly 32%. Across the tested reasoning settings, the error rate stays within 1.9 percentage points of GPT-6 Astra‘s at less than one-fifth the cost per task. This measurement uses deliberately difficult, de-identified ChatGPT conversations where users flagged an earlier model’s error, and the source notes these prompts are not representative of typical usage, where factual errors are rarer.

On safety, GPT-6.1 Sol shows lower failure rates than GPT-6 Sol in challenging evaluations covering transparency about broken search tools, respecting explicit restrictions, and avoiding unauthorized outcomes during agentic tasks, with no observed attempts to bypass an automated safety reviewer. Details are in the GPT-6.1 Sol system card addendum.

GPT-6.1 Sol is available starting today to Plus, Pro, Business, Enterprise, and Edu users in ChatGPT Work and Codex, but not yet in Chat. Developers can access it through the OpenAI API as gpt-6.1-sol at $2 per million input tokens, $0.10 per million cached input tokens, and $10 per million output tokens. A GPT-6.1 Sol Ultrafast variant with up to 8x faster token generation compared with its standard speed in Codex will arrive in the coming days.