OpenAI announced the release of GPT-6.1 Sol on Tuesday, an upgrade to its GPT-6 Sol model, which it had introduced earlier this month. The company stated that GPT-6.1 Sol delivers capabilities approaching those of GPT-6 Astra, its most powerful model, across various complex professional tasks. These tasks include writing and debugging code, understanding documents, and executing multi-step business workflows.
The primary distinction of GPT-6.1 Sol lies in its pricing structure. OpenAI reported that the new model costs one-fifth of GPT-6 Astra's standard input and output token prices. Specifically, GPT-6.1 Sol's standard API prices are $2 per million input tokens and $10 per million output tokens. Cached input tokens are priced at $0.10 per million, which is a 95% reduction from standard input pricing and 50% less than GPT-6 Sol's cached input rate. This pricing adjustment is particularly relevant for developers building agents that frequently reuse context across requests.
OpenAI provided benchmark results to illustrate GPT-6.1 Sol's performance. On DeepSWE v1.1, a benchmark for complex software engineering tasks, GPT-6.1 Sol reportedly matched GPT-6 Astra while costing approximately one-fifth as much. It also surpassed GPT-6 Sol's best score on this benchmark by 6.4 percentage points. For tasks involving computer interaction, evaluated on OSWorld 2.0's offline set, GPT-6.1 Sol outperformed GPT-6 Sol by seven percentage points at maximum reasoning effort and came within 2.1 percentage points of Astra's score, at about one-seventh of Astra's cost per task.
In scientific research, assessed on Terminal-Bench Science 0.1, GPT-6.1 Sol more than doubled GPT-6 Sol's score at maximum reasoning effort, at less than half the cost per task. The average cost per task for GPT-6.1 Sol on this benchmark was $5.47, compared to $23.21 for Anthropic's Opus 5.5 and $23.80 for GPT-6 Astra. The company also noted improvements in factual accuracy, with GPT-6.1 Sol reducing the share of responses containing a factual error from 11.4% to 7.7% at low reasoning effort, representing a 32% reduction compared to GPT-6 Sol.
GPT-6.1 Sol is available to Plus, Pro, Business, Enterprise, and Edu users in ChatGPT Work and Codex, and developers can access it through the OpenAI API. OpenAI also announced plans to offer GPT-6.1 Sol Ultrafast, a faster version of the model, in the coming days, which will provide up to eight times faster token generation in Codex. While GPT-6 Astra remains OpenAI's most capable model for the most demanding projects, GPT-6.1 Sol is positioned to offer a balance of capability and cost for everyday professional tasks and large-scale deployments.