FULL STORY

ChatGPT Voice Gets Model Choice, New Usage Caps

On Sept 10, OpenAI let voice users pick any model and reasoning effort, launching GPT-6 Astra for voice, while simultaneously resetting subscription usage caps with daily voice limits.

2026-09-10 ~ 2026-09-10 · 2 episodes · 13 posts

Episode 1 · ChatGPT Voice Now Supports Model Selection, Including GPT-6 Astra (2026-09-10, 8 posts)

OpenAI employee athyuttamre announced on 09-10 that ChatGPT's voice mode has been upgraded: users can now freely choose any model and reasoning effort in voice conversations, just like in text chats. Pro users can select GPT-5.6 Sol or GPT-5.6 Sol/GPT-6 Astra; voice mode automatically invokes the chosen model when search or harder problems are needed.

Confirmed

  • athyuttamre announced in multiple posts that Voice now supports more models: GPT-5.6 Sol or GPT-6 Astra (Pro users), with freely selectable reasoning effort and automatic invocation when search or reasoning is required
  • Community user ChrisGPT independently discovered that Astra had arrived in ChatGPT Voice, confirming that voice mode can call GPT-5.6 or GPT-6 Astra when search or hard problems come up, with an experience consistent with model selection in text chats
  • Some posts mention increased Plus/Pro limits (from athyuttamre's official announcement)

Why it matters

  • Voice mode has moved from a fixed model to user-selectable models and reasoning effort, closing the capability gap between voice and text chats
  • GPT-6 Astra entering voice mode marks the model's expansion from text to voice interaction scenarios

Episode 2 · OpenAI raises ChatGPT usage limits and adds daily Voice caps (2026-09-10, 5 posts)

On September 10, OpenAI staff member Athyuttam Rele posted a series of updates announcing adjustments to ChatGPT usage caps across subscription tiers: limits for Plus and the Pro $100 plan have been raised, with full details available on the official usage limits update page. The same day, OpenAI engineer Juberti and the news account testingcatalog revealed new caps for voice mode.

Confirmed

  • Usage allowances increased for Plus and Pro $100 tiers (@athyuttamre).
  • Voice mode now has daily limits: Plus gets 3 hours per day (Live model), and the Go tier also gets 3 hours (based on the Live mini model); Pro $100 gets 15 hours; Pro $200 users get unlimited daily usage (@koltregaskes, @testingcatalog).
  • Under the new cap system, voice mode responses slow down when paired with the 5.6 Extra High or 6 Pro models, but this trades off for the highest intelligence in hands-free use (@athyuttamre).
  • OpenAI has released a link to the updated usage limits documentation so users can review the details for each tier.

Why it matters

  • A direct boost in allowance for heavy-paying users, especially the Pro $100 tier, which gets both a general quota increase and a clear voice-time guarantee.
  • Voice mode now has explicit daily time thresholds for the first time (Plus 3 hours, Pro $100 15 hours, Pro $200 unlimited), signaling that OpenAI is moving toward more granular, tiered management of high-frequency voice usage.
  • The "slower but smarter" design approach shows that in hands-free voice scenarios OpenAI prioritizes model intelligence over response speed, requiring users to make a trade-off in experience.