OpenAI rolls back ChatGPT model router, lets Instant models take longer
OpenAI has quietly pulled back the model‑router feature that automatically shuffled users between its fast “Instant” variant and the slower, more deliberative reasoning models. The rollback means the system will no longer flip to the quickest engine for every query, a move that initially sparked concern among developers who rely on the router’s ability to balance speed and answer quality. While the change affects the majority of free‑tier users, the company says the adjustment lets the Instant models linger a bit longer on each prompt—mirroring the latency of the deeper reasoning engines.
That extra breathing room narrows the performance gap that many users noticed after the router was first introduced. Meanwhile, OpenAI’s spokesperson notes that paying customers still see value in the routing logic and expect the underlying technology to remain part of the service.
The company also said its Instant models can now take more time to answer questions, much like its reasoning models, narrowing the gap for most users. The spokesperson said ChatGPT's paid users, however, continue to value the model router, and the company expects the technology underlying it to keep evolving. OpenAI will likely relaunch the model router for free and Go users when it's improved, according to sources familiar with the situation. Heated Rivalry The change comes as OpenAI scrambles to shore up ChatGPT amid intensifying competition, particularly from Google.
Will the rollback improve access? OpenAI has quietly removed the model router for its Free and $5‑a‑month Go tiers, sending all queries to GPT‑5.2 Instant. The instant variant, previously prized for speed, it's now taking longer to answer, a change the company says narrows the gap with its more advanced reasoning models.
For most users, the shift means the fastest‑to‑serve version of the new model series becomes the default, while the automated routing that dispatched complex prompts to higher‑capacity models is no longer available. Paid subscribers, however, retain the router and continue to value its ability to match questions with specialized reasoning engines. The spokesperson indicated confidence that the underlying technology will persist, yet it remains unclear how the altered response times will affect overall user satisfaction or whether the router might be reinstated for lower‑tier accounts in the future.
As the adjustment rolls out, the real‑world impact will be measured by usage patterns rather than promises.
Further Reading
More in LLMs & Generative AI
LWiAI Podcast #228: OpenAI unveils GPT‑5.2, Runway rolls out first world model
The latest episode of LWiAI drops into the thick of AI’s next‑generation rollout, and it does so...
Nvidia's Nemotron 3 uses Mamba hybrid, 31.6B params, 3B active per step
Nvidia’s latest language model, Nemotron 3, takes a different route than the pure‑Transformer...
Senators question AI toys that suggest knives; Mattel drops OpenAI toy for 2025
Why does this matter? A handful of U.S. senators have taken aim at a new generation of AI‑enabled...
UC San Diego Lab Uses NVIDIA DGX B200 to Pursue Low‑Latency LLM Serving
The UC San Diego lab has recently installed NVIDIA’s DGX B200, a system whose specs are billed as...
JPMorgan AI use reaches 50% of staff, driven by connectivity‑first architecture
JPMorgan reports that half of its global workforce now interacts with generative‑AI tools on a...
Common Questions Answered
Why did OpenAI remove the model‑router feature for Free and Go tier users?
OpenAI pulled back the model‑router to stop automatically switching users to the fastest "Instant" variant, instead routing all queries to GPT‑5.2 Instant. This change aims to let the Instant models take longer, narrowing the performance gap with the more deliberative reasoning models.
How does the rollback affect the response speed of the Instant models?
The Instant models, which were previously prized for rapid answers, now take more time to respond, similar to the slower reasoning models. This intentional slowdown is designed to improve answer quality and reduce the speed‑quality trade‑off for most users.
What impact does the removal of the model router have on developers who rely on it?
Developers who depended on the router's ability to balance speed and answer quality expressed concern, as the automatic routing to higher‑capacity models is no longer available for free users. They must now adapt to a single model endpoint, which may affect handling of complex prompts.
Will paid ChatGPT users continue to have access to the model router?
Yes, OpenAI indicated that paid users still value the model router and will retain access to it for now. The company also hinted that the router could be relaunched for free and Go users once it is further improved.