OpenAI slashes GPT‑5.6 Sol API prices, turning up the heat in the AI cost war

OpenAI has quietly cut API prices for its current frontier model, GPT‑5.6 Sol, by as much as one‑third, according to tracking by model‑benchmark site ModelDex. GPT‑5.6 Sol powers advanced chatbots and autonomous AI agents, and its inference costs — the price for each request — have been a major line item for startups and large enterprises alike.

The move comes as the “cost per token” battle between leading AI providers intensifies. For use cases like always‑on AI agents, customer support automation, or long‑form document analysis, even small price differences can translate into millions of dollars in annual cloud spend. Cheaper access to GPT‑5.6 Sol could make it easier for smaller companies to experiment with complex workflows that previously only big tech firms could afford.

At the same time, OpenAI still has to fund safety research, evaluations, and infrastructure for its frontier systems, raising questions about how far prices can fall without squeezing margins. The new pricing underscores a broader shift: frontier‑model quality is converging, so vendors are competing not just on capabilities, but on economics and reliability — a trend likely to accelerate as AI becomes basic infrastructure.