Starting today, we’re introducing three level of intelligence you can choose from. 1) Low: 4x cheaper than the medium mode and use GPT-5.4 mini (high effort). 2) Medium: run using GPT-5.4 3) High: twice as expensive as Medium and use GPT-5.5 (medium effort) frontier model
We're committed to not overwhelming users with choice anxiety over all the models out there (and the effort involved). Instead, Notis automatically picks the right level of intelligence based on the task to always find the best compromise between speed, cost, and intelligence.
We’re also letting you choose the priority between normal, priority (faster at twice the price) and flex (slower at half the price).
If you’re on a PRO+ or ULTRA plan and want a little more control, you can also override the auto mode in the following places:
Per conversation: from the manager overlay and chat, you can choose the thread model and priority mode.
Per channel: From the channel page, you can customize the default model and speed for each of your channel.
Per automation: You can also change the model for each automation (automations always run as flex speed to minimize your cost).
The new overrides are available in the Beta release channel to everyone starting today.
In any case, the Notis apps gives you full visibility on the intelligence used for each request in the timeline of the conversation.
Enes Demirağ
•
Apr 27
I agree. A model agnostic system would be awesome. Users can select between different providers (OpenAI, Antropic, Google) and different intelligence levels (small or large models)
With this way, If a user wants they can trade-off between a dummer model and lower costs.
Log in to comment and vote
Comments2
Florian (Flo) Pariset
Jun 11
Starting today, we’re introducing three level of intelligence you can choose from.
1) Low: 4x cheaper than the medium mode and use GPT-5.4 mini (high effort).
2) Medium: run using GPT-5.4
3) High: twice as expensive as Medium and use GPT-5.5 (medium effort) frontier model
We're committed to not overwhelming users with choice anxiety over all the models out there (and the effort involved). Instead, Notis automatically picks the right level of intelligence based on the task to always find the best compromise between speed, cost, and intelligence.
We’re also letting you choose the priority between normal, priority (faster at twice the price) and flex (slower at half the price).
If you’re on a PRO+ or ULTRA plan and want a little more control, you can also override the auto mode in the following places:
Per conversation: from the manager overlay and chat, you can choose the thread model and priority mode.
Per channel: From the channel page, you can customize the default model and speed for each of your channel.
Per automation: You can also change the model for each automation (automations always run as flex speed to minimize your cost).
The new overrides are available in the Beta release channel to everyone starting today.
In any case, the Notis apps gives you full visibility on the intelligence used for each request in the timeline of the conversation.
Enes Demirağ
Apr 27
I agree. A model agnostic system would be awesome. Users can select between different providers (OpenAI, Antropic, Google) and different intelligence levels (small or large models)
With this way, If a user wants they can trade-off between a dummer model and lower costs.