Manus Free Chat & Gemini Updates | Enhanced AI Experience

Manus has added a chat without credit limits, along with 1.5 interesting news items that I haven’t shared before but are worth knowing.

About Manus

Generally (this is a personal assumption — not confirmed) — they realized that an agent, heavily reliant on credits, doesn’t stay long with users within the interface. This likely leads to low LTV. Therefore, the current focus is on keeping users engaged on the platform longer to get them used to the tool. That’s why the model was made free.

Overall, it has become significantly more convenient. Now, simple tasks can be handled with the free tool, while more complex ones can be delegated to the agent.

They also added Gemini. The chat mentions that it’s not just a single model, but a set of solutions aimed at improving the user experience, so it doesn’t give definitive answers to questions.

Main advantages:
– It can be used without deducting credits, which is very convenient.
– It understands documents and files well.
– Quick switch to an agent with one click, which is also very handy.

To activate the chat, you need to click the button shown in the screenshot; otherwise, the agent starts by default, and credits will be used up.

Here are the links:
– to Manus
– to my experience working with it

Now, moving to news I haven’t publicized before.

OpenAI launched a new model — o3-pro. Currently, it’s only available to Pro accounts (costs around $200) and via API. The model is truly impressive — deserving top marks. It’s smart, beautiful, and internally complex.

The most interesting part — it’s significantly cheaper than o1 Pro. And quite substantially, making it accessible to a much larger audience — likes and subscriptions are welcome.

And to conclude — I’ve now also set the notification bell. The o3 model, which integrates well with deep research and can do a lot, has also become noticeably cheaper — roughly five times less, based on my estimates.

Additionally, query limits for regular users have been increased compared to before. Previously, it was 50, then raised to 100, and now users are given as many as 200 requests per week. More details can be found through the links.

And a bonus — 0.25 more news items. Mistral has finally released its own model. However, it’s still very raw and doesn’t attract much interest yet. Nonetheless, I’ll keep you posted.

Here is the link.

Postscript: if you get blocked when posting a comment, it’s probably because you don’t have a username set in Telegram. If you do, we recommend reading this material to understand the reason.

Page view /ai-blog/manus-free-chat-gemini-updates-enhanced-ai-experience/ 01.05 05:51 Page view /ai-blog/china-holds-large-scale-live-fire-drills-in-taiwan-strait-global-military-exercises/ 01.05 05:40 Page view /ai-blog/google-flow-worldwide-access-no-vpn-needed-google-labs/ 01.05 05:15 Page view /ai-blog/kimi-2-5-vs-composer-2-licensing-model-controversy-explored/ 01.05 04:54 Page view 01.05 04:36 Page view /ai-blog/trump-releases-greenland-correspondence-macron-nato-insights/ 01.05 04:33 Page view /ai-blog/upcoming-nvidia-rubin-vera-chips-faster-launches-in-2023/ 01.05 04:29 Page view 01.05 04:27 Page view 01.05 04:22 Page view /category/ai-blog/ 01.05 04:22