Gemini 3,5 is expected in March 2026.

At Google IO Event 2026.

yo! honestly i get the struggle with those context windows cuz i'm basically doing the same thing with heavy coding lately!! i've been playing around with [[PRODUCT:Google Gemini 1.5 Pro]] and it's amazing, but yeah, hitting limits is mood.

sooo about Gemini 3.5... there's literally no official date yet. i think we might actually see a [[PRODUCT:Google Gemini 2.0]] before 3.5 even exists? Google's cycle is usually like every 6-8 months for big jumps.

Option A: Commit now to [[PRODUCT:Google Gemini Enterprise]] for the high-tier support.
Option B: Wait for the next I/O event whispers.

Pros for A: You get better throughput NOW. Cons: You might feel FOMO if a new model drops in Oct. Personally? I'd just stick with the [[PRODUCT:Google Gemini 1.5 Flash]] for quick tasks and wait a bit longer to see if the 2.0 rumors are real. i mean, the jump from 1.0 to 1.5 was huge, so 2.0 or 3.0 will be FANTASTIC whenever it lands! gl!

TL;DR: No leaked date for 3.5 yet, probably 2.0 first later this year. wait if ur unsure!

Similar situation here - I went through this exact same dilemma a few months back when trying to scale my dev team's output. Honestly, I was highkey stressed about committing to a massive contract when everything is changing so fast. I ended up sticking with the Google Cloud Vertex AI pay-as-you-go model instead of a fixed enterprise sub, and honestly, it saved us a ton.

I mean, Gemini 3.5 rumors are all over the place, but based on how Google dropped previous versions, I wouldn't bank on a specific date yet. To manage costs while waiting, I've been doing a few things:
* Using DeepSeek Coder V2 for the simpler stuff cuz it's super cheap
* Running certain tasks through the OpenRouter API to compare outputs
* Sticking to the free tier of GitHub Copilot Individual for basic autocomplete to save tokens

It's kinda a gamble, but staying flexible might be better than locking in now. Are you noticing the reasoning degrades when the context gets full, or is it just cutting you off? lol the struggle is real...

Seconded!

> I’m currently using Gemini for heavy coding workflows and complex data analysis, but I’m hitting some context window limitations.

Curious about one thing: are you actually maxing out the 2M tokens on the current Google Gemini 1.5 Pro API or just the web version? tbh if you're hitting limits there, you might wanna check out Anthropic Claude 3.5 Sonnet for coding while we wait, but I'd love to know your specific file sizes first! lol

> I’m currently using Gemini for heavy coding workflows and complex data analysis, but I’m hitting some context window limitations.

Curious about one thing: are you actually maxing out the 2M tokens on the current Google Gemini 1.5 Pro API or just the web version? tbh if you're hitting limits there, you might wanna check out Anthropic Claude 3.5 Sonnet for coding while we wait, but I'd love to know your specific file sizes first! lol

Did this last week, worked perfectly

> I ended up sticking with the Google Cloud Vertex AI pay-as-you-go model instead of a fixed enterprise sub, and honestly, it saved us a ton. Totally agree with going the pay-as-you-go route for now. Committing to a heavy contract when a big jump like a Gemini 3.5 or even a 2.0 could drop any day is super risky. If youre struggling with reasoning or limits while waiting for the next Google update, here is how a few other stacks compare for those heavy dev workflows: 1. OpenAI GPT-4o: Pros: Generally tighter reasoning on complex logic and better at following strict system instructions. Cons: The context window is tiny compared to Geminis 2M tokens, so youll definately need a solid RAG pipeline or some serious chunking.
2. Microsoft Azure OpenAI Service: Pros: You get the GPT models but with better privacy controls if youre worried about enterprise data. Cons: Provisioning throughput can be a real pain compared to how easy Google makes it. I mean, rumors are floating that the next big release might focus on "agentic" workflows rather than just expanding the window. Wait no, basically what I'm saying is that the next leap might be more about *how* the model uses the data than just how much it can hold. Tbh, I'd stay flexible until we see the official roadmap.

Honestly, nobody actually knows the date for 3.5 yet and it drives me crazy how these tech giants keep everyone in the dark while dangling new versions over our heads. It feels like a total money grab to get people to sign these massive enterprise contracts right before a jump happens. The pricing is honestly ridiculous for what you get half the time... like, we are basically paying a premium to be beta testers while they figure out how to stop the models from tripping over basic logic. Its such a scam that the features we actually need always come with some fine print or a price tag that makes no sense for a smaller team. I wouldnt commit to any long-term sub right now because the quality-to-price ratio is just going downhill while they focus on hype cycles instead of actual stability. It is frustrating as hell trying to plan a workflow when the goalposts move every few months and they dont care how much it messes with our budgets.

Building on the earlier suggestion, i am stuck in this exact same loop and it is beyond annoying. I have been digging through every dev blog and leak site for months now trying to find a solid date because my massive data sets are just killing the current context window. It really sucks being a long-time user and having zero visibility on when the next jump is coming... i am totally paralyzed on whether to sign a big enterprise deal or just keep limping along with what we have for now.

Seconded!