The artificial intelligence landscape shifts once again as AI/ML API announces the immediate availability of OpenAI's latest flagship offering, GPT-6.1 Sol. Bringing near-Astra performance to complex coding and agentic workflows at a fraction of the traditional cost, this release marks a significant milestone in how builders and enterprises deploy reasoning models at scale.
According to data from AI/ML API and Artificial Analysis, GPT-6.1 Sol bridges the gap between high-end reasoning capabilities and cost efficiency. While high-performance models like GPT-6 Astra have dominated advanced agentic workflows and computer use, their steep pricing of $13 per million input tokens and $65 per million output tokens has created budget friction for scaling businesses. GPT-6.1 Sol changes this equation by delivering comparable professional-grade results at roughly one-fifth of Astra's token price, maintaining the same $2.60 per million input and $13 per million output pricing structure as its predecessor while drastically elevating quality.
Under the hood, GPT-6.1 Sol boasts a massive context window of 1,050,000 tokens with up to 128,000 tokens of output capacity. The model integrates a suite of native features designed for autonomous systems, including built-in tool calling, advanced web search, file search, structured outputs, and configurable reasoning with effort levels ranging from low to max. Prompt caching is also fully supported, further reducing overhead with cached input priced at just $0.13 per million tokens.
For founders and engineering leaders, adopting GPT-6.1 Sol requires minimal engineering friction. Because the model is fully OpenAI-compatible, engineering teams can integrate it simply by swapping out the base URL to point at the standard chat completions endpoint and setting the model parameter to openai/gpt-6.1-sol. This drop-in compatibility allows organizations to immediately upgrade their agentic infrastructure without rewriting their existing SDK integrations.
Ultimately, GPT-6.1 Sol democratizes access to state-of-the-art agentic coding and reasoning. By pairing a million-token context with aggressive cost optimization, OpenAI and AI/ML API are giving builders the headroom required to run complex, long-running autonomous workflows without compromising on financial sustainability.