Vishal Dharmadhikari

@bernaferrari @reah_ai If you're using AI Studio, you can enable Logs for your project. This allows you to track all the requests made via the project ai.google.dev/gemini-api/doc…

@reah_ai I wish you added a long random ID to every Gemini API generation so that I can open the web and see the request (what did it error, what google server returned). Sometimes there are errors and my api didn't catch and I'll never see them again.

two months ago, I (re)joined Google! as Senior Director, Google Models, my team's mission is to bridge frontier intelligence for the enterprise and transform work. simply put, shipping industry leading models built for real world use cases. the days of clapping for hillclimbing academic benchmarks are behind us :) DMs open for feedback on Gemini, NanoBanana, Veo3, Live, Lyria, AlphaGenome, and all other models.

@mathlover3b1 @rekram11 @LukeParkerDev Would you be able to send me a DM? Wasn't able to send you one.

@rekram11 @Vish_owl @LukeParkerDev same question :)

gemini 3.1 literally just uses bash to write files to use node to edit other files its so over

@rekram11 @LukeParkerDev Sent you a DM!

@Vish_owl @LukeParkerDev Do you have that model available on google vertex yet? @Vish_owl

@MaximeRivest Sorry about that - we're working on stabilizing things

@Vish_owl thanks! api is a bit slow but i expect so early in the preview/release!

When gemini 3 came out I was super excited because of its performance on these benchmarks. I have come to believe that gemini is the best general model but we have a harness problem. I think I will try to get `pi` in a shape that uses gemini 'optimally'.

@benediktstroebl @_philschmid Benchmarks & performance was neutral so you shouldn't see any quality impact from the change, just better tool selection

@Vish_owl @_philschmid thanks! Do you have any evals on this?

@_philschmid Can you share some more details on `gemini-3.1-pro-preview-customtools`

@LukeParkerDev Benchmarks & performance was neutral so you shouldn't see any quality impact from the change, just better tool selection

@Vish_owl Watching it use bash and script things is kinda fun anyway. But yeah using tool calling will be awesome once we get it working. I'm curious - any differences in benchmarks/performance when using that endpoint?

@CallMeMrYogurt Hey! Was bash the correct decision here, or would you expect it to use another customer tool you've defined?

Gemini 3.1 Pro is the first model that uses bash to add go dependecys not create files with random versions #google #ai #gemini

@polymorfaisal @noah_vandal @kimmonismus If you test out this endpoint, we'd love to get your feedback!

@noah_vandal @kimmonismus gemini-3.1-pro-preview-customtools * For those building with a mix of bash and custom tools, Gemini 3.1 Pro Preview comes with a separate endpoint available via the API called. This endpoint is better at prioritizing your custom tools (for example view_file or search_code).

Gemini 3.0 pro preview was released Nov 18, 2025 Within 3 months Google achieved significant jumps in performance. 3.1 is not only (on paper) SOTA on most benchmarks, I find it impressive that Google is only assuming an iterative improvement in the model number for this big leap. Previously this would have been a jump to 3.5.

@mathlover3b1 @LukeParkerDev This may be helpful: x.com/Vish_owl/statu…

@LukeParkerDev it is still fucking ass at toolcalling it's joever looks intelligent tho, but man it keeps spamming edits before reading files, and uses cat on the whole file instead of using the read tool, etc etc etc... i still have my heart in google cuz this model is old tho

@ammaar Will you teach us taste?

The Google AI Professional Certificate is live on Coursera and it’s free! So excited to be teaching in this. With Gemini 3.1 Pro, I’ll show you how to vibe code apps + use AI to give you a massive edge at work. And you get 3 free months of Google AI Pro. See you in class! :)

@osanseviero Always has been

It still is

Excited for ARC-AGI-3!

@alex__bod @GoogleAIStudio Hey @alex__bod, Google AI Studio & Gemini API access in Andorra is now live!

@googleaistudio Can you make Google AI Studio available in Andorra?

🍌 nano banana is here → gemini-2.5-flash-image-preview - SOTA image generation and editing - incredible character consistency - lightning fast available in preview in AI Studio and the Gemini API

@IvanPorozhnyy @GoogleAI @googledevs @OfficialLoganK Hey @IvanPorozhnyy, access to AI Studio and the Gemini API in Moldova is now live!

Hey @GoogleAI @GoogleDevs @OfficialLoganK Developers in Moldova🇲🇩already have access to Google AI Pro and AI Ultra. However, AI Studio & API remain geoblocked. There are multiple requests from our community about this on discuss.ai.google.dev Time to unlock the region? 🚀

@VladislavNegura @OfficialLoganK Access to AI Studio and the Gemini API in Moldova is now live! Happy building

@OfficialLoganK can we finally enable gemini api for Moldova?

time to start shipping again : )

@OfficialLoganK @matthewmarley @_philschmid @GoogleAIStudio This is live now! x.com/GoogleAIStudio…

@matthewmarley @_philschmid @GoogleAIStudio coming soon! We will let you use GCS storage

You haven’t tried @GoogleAIStudio yet?👀 We made it simpler! When you come to AIS for the first time, you will have a Default Gemini Project & API Key waiting for you! This should reduce time to first prompt, and help you start building faster! Give it a try!

@harrisonfjobe @GoogleAIStudio Tis the season for checklists after all

ohhhh new active checkmarks in @GoogleAIStudio playground