Gemini Flash 2.5 Image - now on Discourse

[sam (Sam Saffron)]

Google just released an exciting new multi-model LLM.

Image editing in Gemini just got a major upgrade

This model is fast and extremely powerful.

This change was introduced in the following commit:

https://github.com/discourse/discourse/commit/00c98781c19378725865b6beddfbbf266cb09dde

I made a quick video demonstrating the feature:

Gemini Nano Banana and Discourse

See previously: OpenAI image generation and editing now supported in Discourse AI

AI Personas are documented at: AI bot - Agents

Configuration in a nutshell

• Add Gemini Flash 2.5 Image
• Enable vision and bot
• Enable vision on your creative persona
• Make pictures!

[merefield (Robert)]

All our group pictures in Google cloud storage going to good use, clearly!

The plinth was a nice touch!

[sam (Sam Saffron)]

Nano Banana 3 is out and supported in Discourse.

Amazing how it transported us to Tuscany, Robin and Jeff look way off, but the lighting and definition and background is all pretty spectacular.

[NateDhaliwal]

Now that… that is amazing. I find it really nice that there’s less image inaccuracies/inconsistencies now. Time to add and use this!