OpenAI adds GPT-5.1 to its API, featuring a “no reasoning” mode and extended prompt caching with up to 24-hour retention for faster responses at a lower cost
Today we're releasing GPT‑5.1 in the API platform, the next model in the GPT‑5 series that balances intelligence …
OpenAI
Context & Ripple Effects
OpenAI’s API story began with GPT-4’s advanced-reasoning API rollout and later expanded into more explicit model variants and larger context options. GPT-5.1 makes serving behavior—not only raw capability—a product choice for API builders.
API customers can choose a no-reasoning path when response speed and cost matter more than deliberative processing.
Applications that reuse large prompt prefixes can retain cached prompts for up to 24 hours, reducing repeated work and, according to OpenAI, lowering latency and cost.
Second-order effects
Developers may segment workloads by task, reserving reasoning-enabled calls for harder requests while routing routine, repeatable interactions to the faster mode.
Extended cache retention makes prompt architecture and request timing more consequential for application costs, increasing pressure on competing API platforms to offer comparable serving controls.
Third-order effects
Frontier-model APIs are increasingly differentiated by controllable inference economics—latency, caching, and reasoning depth—rather than a single headline capability measure.
If this pattern persists, application teams will design around provider-specific routing and cache semantics, making operational integration a stronger source of platform lock-in even as model choices multiply.
The trend: AI model platforms are turning inference-time controls into core product tiers, letting customers trade reasoning depth against speed and cost by workload.
@OpenAIDevs This model is amazing! GPT-5.1 would often take 100+ seconds to use my call for simple, multi-step function calls and then give poor results. Now it's 9 seconds! Cost per query also gone down by a factor of 10 for these simple queries... Supreme effort OpenAI. Well do…
For coding, you'll see the model is more steerable, communicative, and produces higher-quality code. It overthinks less, fixes more, and stays on target during long tool-call sequences. [image]
We also shipped gpt-5.1-codex and gpt-5.1-codex-mini. Here's how to think about these models: • gpt-5.1 for everyday coding tasks • gpt-5.1-codex for complex, long-running agentic coding • gpt-5.1-codex-mini for cost-efficient edits and changes
GPT-5.1 adapts its reasoning: fewer tokens for simple tasks, deeper thinking for complex ones. We're also introducing “no reasoning” mode so it responds faster on tasks that don't need deep thinking. It's the default reasoning_effort setting. [image]
GPT-5.1 is now available in the API. Pricing is the same as GPT-5. We are also releasing gpt-5.1-codex and gpt-5.1-codex-mini in the API, specialized for long-running coding tasks. Prompt caching now lasts up to 24 hours! Updated evals in our blog post.