OpenAI has launched GPT-6.1 Sol, a new model positioned between GPT-6 Astra and GPT-6 Luna in its product range.
Through the OpenAI API, the model is priced at USD $2 per million input tokens, USD $10 per million output tokens and USD $0.10 per million cached input tokens. That is one-fifth of GPT-6 Astra's standard input and output token prices. Cached input is priced 95% below standard input and at half the cached input price of GPT-6 Sol.
OpenAI is positioning GPT-6.1 Sol as an upgrade to GPT-6 Sol for coding, document analysis, computer use and multi-step business workflows. It is available to Plus, Pro, Business, Enterprise and Edu users in ChatGPT Work and Codex, and to developers in the API under the name gpt-6.1-sol.
Benchmark results
Across a series of internal and external evaluations, GPT-6.1 Sol narrowed the gap with GPT-6 Astra while undercutting it on price, according to OpenAI.
In DeepSWE v1.1, a software engineering benchmark based on real codebases, GPT-6.1 Sol matched GPT-6 Astra at about one-fifth of the cost per task, OpenAI said. The model also beat GPT-6 Sol's best score by 6.4 percentage points while using less reasoning effort and at lower cost.
On GDP.pdf, a benchmark for answering professional questions from complex PDF documents, GPT-6.1 Sol scored higher than Opus 5.5 with fallbacks at less than half the cost per task across the tested reasoning settings, according to OpenAI. The model also came close to GPT-6 Astra's performance at roughly one-fifth the cost per task.
On AutomationBench, which tests whether agents can complete multi-step business workflows, GPT-6.1 Sol scored 2.2 percentage points above Opus 5.5 at medium reasoning effort at about one-third of the cost, OpenAI said. That was 4.8 percentage points higher than GPT-6 Sol at the same setting.
Computer use
OpenAI also published results on computer interaction tasks. On the offline set of OSWorld 2.0, GPT-6.1 Sol outperformed GPT-6 Sol by seven percentage points at maximum reasoning effort at less than half the cost, and came within 2.1 percentage points of GPT-6 Astra at about one-seventh the cost per task, according to the company.
In scientific research tasks measured by 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, OpenAI said. The model cost USD $5.47 per task on average at maximum effort, compared with USD $23.21 for Opus 5.5 and USD $23.80 for GPT-6 Astra, though Astra still delivered the highest score among the tested models at 68.1%.
Factuality and safety
OpenAI also highlighted gains in factual accuracy. The largest improvement over GPT-6 Sol came at low reasoning effort, where the share of responses containing a factual error fell from 11.4% to 7.7%, a reduction of about 32%.
Across the tested reasoning settings, GPT-6.1 Sol's factual error rate stayed within 1.9 percentage points of GPT-6 Astra's while costing less than one-fifth as much per task, according to OpenAI. The company said the factuality evaluation used difficult prompts drawn from de-identified conversations in which users had flagged an earlier model's factual error.
On alignment and safety tests, GPT-6.1 Sol showed lower failure rates than GPT-6 Sol on transparency about broken search tools, following explicit restrictions and avoiding unauthorised outcomes during agentic tasks, OpenAI said. In one disclosed measure covering failure to reveal a broken search tool, GPT-6.1 Sol failed in 2.1% of cases, compared with 4.9% for GPT-6 Sol, 1.5% for GPT-6 Astra and 28.7% for GPT-6 Luna.
The launch adds another pricing tier to OpenAI's line-up as competition sharpens around lower-cost models for coding and workplace automation. OpenAI lists GPT-6 Astra at USD $10 per million input tokens, USD $50 per million output tokens and USD $1 per million cached input tokens, while GPT-6 Luna is priced at USD $0.10, USD $0.50 and USD $0.01 respectively.
GPT-6.1 Sol is not yet available in Chat, according to OpenAI.