Google limits free Gemini app users to Flash-Lite from Oct. 9; AI Pro adds Deep Think
Starting October 9, Gemini app users without a Google AI subscription will be limited to Flash-Lite, losing access to Flash and Pro. AI Plus ($4.99/month) subscribers lose Pro, while AI Pro ($19.99/month) gains Deep Think, according to an updated Google support document.
Google will restrict which models free and AI Plus users can access in the Gemini app, with the first changes starting October 9, according to an updated Google support document reported by 9to5Google. Users without a Google AI subscription will be limited to Flash-Lite. Meanwhile, the $19.99/month AI Pro plan gains Deep Think.
What changes, by plan
Per the support document, as reported by 9to5Google, the model versions involved are Flash-Lite (3.5), Flash (3.6) and Pro (3.1). Access will break down as follows:
| Google AI plan | Flash-Lite | Flash | Pro |
|---|---|---|---|
| No plan (free) | Yes | No | No |
| AI Plus ($4.99/mo) | Yes | Yes | No |
| AI Pro ($19.99/mo) | Yes | Yes | Yes |
| AI Ultra | Yes | Yes | Yes |
- Free users: From October 9, users who do not pay for Gemini lose access to Flash (3.6) and Pro (3.1).
- AI Plus subscribers: They keep Flash-Lite and Flash but lose Pro. Google says these subscribers will "receive an email that explains when these changes will take effect." A specific date for AI Plus was not given in the source.
- AI Pro and AI Ultra: There are no regressions for either tier.
Deep Think comes to AI Pro
AI Pro subscribers will get access to the Deep Think option, which Google describes as "maximum parallel reasoning." It is currently available only on the $99.99/month and $199.99/month plans. The source does not state an exact rollout date for Deep Think on AI Pro.
New effort levels
Google also appears to be adding "low," "medium," and "high" effort levels "for each available model" in October. Higher levels increase the model's ability to complete tasks and give more thorough answers, but they consume more of the user's usage limit. Today, the app offers a single option, "Extended thinking: Complex problem solving." The three tiers would match what is already available in Google AI Studio and Antigravity.
Context
The changes follow a compute-based usage change Google introduced in May. Google has not disclosed context window sizes or per-token pricing as part of this update. Those figures are not relevant to the consumer app plans, and the support document as reported does not include them. The details come from a support page rather than a formal announcement, so timing and specifics may change.
What this means
Google is shifting from limiting how much users can use Gemini to limiting which models they can use. Free and AI Plus users now sit at the bottom of a clearer capability ladder, while the $19.99 AI Pro plan becomes the entry point for both Pro-tier models and Deep Think. That pushes users who want top-tier reasoning toward a paid tier.
The effort levels point the same way. If higher effort draws down a user's limit faster, Google is giving subscribers a direct way to trade compute for quality, and making that trade-off visible. Matching AI Studio and Antigravity also gives developers and consumers a more consistent control across Google's products.
The $4.99 AI Plus tier takes the biggest hit relative to expectations, since it loses Pro entirely. Users on that plan who rely on Pro will need to decide whether to move to AI Pro or accept Flash. Free users, who lose Flash as well, will see the largest drop in capability.
Related Articles
Google launches Playground, a free text-to-game platform for US adults built on Gemini, Nano Banana and Lyria
Google has launched Playground, a browser-based platform that lets US adults create games from text prompts alone. It runs on Gemini, Nano Banana and Lyria, according to a Google spokesperson, and is free with higher weekly limits for Google One subscribers. Google and Unity also announced Unity Spark, a closed-beta tool for professional developers.
Google releases free AI Edge Foresight, a Mac note-taking app that transcribes meetings fully offline
Google has released AI Edge Foresight, a free experimental macOS app that transcribes meetings and audio files entirely on-device. It is optimized for Apple Silicon Macs and uses Google's EmbeddingGemma 2 model, according to The Verge. Google says audio, files, and notes never leave the user's computer.
Google Cloud launches Gemini agent, a multi-model enterprise agent that runs Gemini and Claude, in private preview
Google Cloud announced the Gemini agent at Gemini at Work 2026 on October 8. It is an enterprise agent that routes each job to either Gemini or Claude models. It is in private preview, with wider availability for select Workspace Business and Enterprise plans coming soon.
Google's SynthID detector goes global in English; Apple Intelligence watermarking to follow 'soon'
Google has moved its SynthID AI content detector from limited early access to global availability in English. The tool now covers content from OpenAI, Nvidia, Kakao and, "soon," Apple Intelligence, in addition to Google's own Gemini models.
Comments
Loading...