GPT-5.6 Sol gets a ChatGPT-specific behavior update, not a new Work or Codex model

OpenAI says that Plus and Pro users are receiving an updated GPT-5.6 Sol in ChatGPT. The stated changes target more focused answers, better use of factual sources, and a new slider that lets a user choose how much thought ChatGPT applies to a response. OpenAI also says the same ChatGPT model now serves both Instant responses and deeper reasoning, which is intended to make a higher-effort response feel like a continuation of the same assistant rather than a switch to a differently behaved model.

The timing matters because OpenAI had described GPT-5.6 Sol's API changes only a week earlier. On July 30, it introduced Fast mode for Sol, said it can deliver up to 2.5 times the speed of Standard processing at twice the price, and kept Sol pricing unchanged. The August 6 announcement is instead a product-behavior update inside ChatGPT. It pairs the Sol change with a separate access change: GPT-5.6 Luna becomes the default for Free and Go users, with unlimited text chats and a Think button scheduled to roll out subject to abuse guardrails.

The strongest practical distinction is between response policy and model availability. OpenAI reports that, in an internal evaluation of financial, medical, and legal prompts requiring factual detail, answers with at least one factual error were about 68% less common for GPT-5.6 Sol than for GPT-5.5 Instant. That is useful directional evidence, but it is an OpenAI evaluation rather than an independently reproduced result. The company also frames the change around concise answers and fewer unnecessary formatting choices, so teams should evaluate whether those preferences improve their own prompts and workflows rather than treating the examples as a universal quality measure.

For developers and organizations using Work or Codex, the release has a clear limitation: OpenAI explicitly says that the GPT-5.6 Sol version powering those products is not changing. The new ChatGPT slider therefore should not be assumed to expose an equivalent API, Codex, or Work control. Technical users should separate ChatGPT interaction testing from production evaluation, especially where latency, tool use, output formatting, or model routing matters. OpenAI's own July description suggests Fast mode remains the relevant Sol lever for API workloads where reduced response time justifies a higher processing price.