Logan Kilpatrick from Google DeepMind explains the accessibility and capabilities of Google's Gemini 3 Pro model to host Greg Isenberg.
Assertion Partly supported
Kilpatrick: Gemini 3.0 Pro API undercuts GPT-5.1 Pro and Claude 4.5
“The actual model is two dollars per million input tokens and 12 dollars per million output tokens. So it's priced below some of those sort of comparable competitive models like GBT, 5.1 pro as an example, and Claude 4.5.”
Insight
Direct AI creation preserves story fidelity better than human intermediary layers
“What happens when you, as the person who like has a story to tell or like some idea, you have to like put it through this translate, this translation layer, especially when the translation layer is sometimes like a, you know, a bunch of people or something lik…”
Assertion Contradicted
Image quality does not degrade across multi-turn AI conversational edits
“The model doesn't like the image quality doesn't get worse as you like do a multi-turn edit. So like you don't need to get everything right on the first edit.”
Prediction Not checkable as stated
Kilpatrick: Real-time multimodal models will become co-present pair programmers
“I think the future that we're going to see with the multimodal live API is really the ability for the model to or just like generally for models to be co-present in the experiences that you are. And like, you could imagine your IDE is the best experience of th…”
Assertion Not checkable as stated
Kilpatrick: Users can vibe-code real-time games with Gemini 3 Pro
“And I think with Gemini Three Pro, you can just, Vibe code games in real time, which is really, really awesome.”
Insight
Kilpatrick: Generating games is a proxy for an LLM's overall sophistication
“Games are just a proxy for like the model's overall capability. So like, Hey, maybe you're like, I don't actually care about building games. I think as you see some of these examples, it like speaks to the level of sophistication that the model has an understa…”