---
title: "Gemini 3.7 Flash and Grok 4.6 in Chat Thing"
canonical_url: "https://chatthing.ai/blog/gemini-3-7-flash-and-grok-4-6"
last_updated: "2026-08-19T12:26:08.884Z"
meta:
description: "Try Gemini 3.7 Flash and Grok 4.6 in Chat Thing, with vision, Power-Ups and large context windows. Plus clearer Power-Up activity and Google Drive imports."
"og:description": "Try Gemini 3.7 Flash and Grok 4.6 in Chat Thing, with vision, Power-Ups and large context windows. Plus clearer Power-Up activity and Google Drive imports."
"og:title": "Gemini 3.7 Flash and Grok 4.6 land in Chat Thing - Blog - Chat Thing"
"twitter:description": "Try Gemini 3.7 Flash and Grok 4.6 in Chat Thing, with vision, Power-Ups and large context windows. Plus clearer Power-Up activity and Google Drive imports."
"twitter:title": "Gemini 3.7 Flash and Grok 4.6 land in Chat Thing - Blog - Chat Thing"
---
[← Back to blog](https://chatthing.ai/blog)**Changelog**# **Gemini 3.7 Flash and Grok 4.6 land in Chat Thing**
Zef
19 Aug 2026
~ 4 min read
---
---
### **On this page**
1. Gemini 3.7 Flash: the practical first pick
2. Grok 4.6: a more deliberate choice
3. Gemini 3.7 Flash vs Grok 4.6
4. [Power-Up activity, without all the clutter](#power-up-activity-without-all-the-clutter)
5. [More file types in Google Drive sync](#more-file-types-in-google-drive-sync)
6. [Try the new models](#try-the-new-models)
Two new models have joined [Chat Thing's model lineup](https://chatthing.ai/models): Google's Gemini 3.7 Flash and xAI's Grok 4.6.
Both can work with images, call [Power-Ups](https://chatthing.ai/pages/features/power-ups) and handle much more context than a typical support conversation needs. They are built for different jobs, though. Gemini 3.7 Flash is the lighter everyday option. Grok 4.6 uses more message tokens, but gives you another model to test when an agent needs more reasoning.
This release also makes Power-Up activity much easier to follow and adds more file types to Google Drive sync.
## Gemini 3.7 Flash: the practical first pick
Gemini 3.7 Flash has a context window of 1,048,576 tokens, supports vision and can use Power-Ups.
It is also the lighter of the two new models on Chat Thing's message-token allowance. That makes it the one we would test first for a busy support agent, especially if it works with a large knowledge base, long documents or image attachments.
The big win is speed. Gemini 3.7 Flash is very fast, which matters when customers are waiting and an agent needs to keep up with a steady flow of conversations.
Benchmarks are useful, but we would not pick a customer-support model from a leaderboard alone. Test whether it answers your questions accurately, follows your instructions and calls the right Power-Ups across the conversations your customers actually have.
## Grok 4.6: a more deliberate choice
Grok 4.6 has a 500,000-token context window, vision and full Power-Up support.
It costs more message tokens than Gemini 3.7 Flash, so it is not the model we would switch every high-volume agent to without testing. Use it to compare a different reasoning model on harder conversations, multi-step tasks or agents that need to decide when and how to use several tools.
xAI positions Grok as a [truth-seeking assistant](https://x.ai/grok). That positioning may suit an agent that should answer more directly and with less buttoned-up prose rather than hedge every point. That tone will not suit every customer-facing agent, so test it against your instructions and escalation rules before switching a live agent.
## Gemini 3.7 Flash vs Grok 4.6
| | Gemini 3.7 Flash | Grok 4.6 |
| --- | --- | --- |
| Context window | 1,048,576 tokens | 500,000 tokens |
| Chat Thing message-token allowance | Lighter | Heavier |
| Where we would test it first | High-volume support, large knowledge bases and image-heavy questions | Harder multi-step tasks, longer Power-Up sequences and agents that need a more direct tone |
[Artificial Analysis publishes high-reasoning results for both models](https://artificialanalysis.ai/models/comparisons/gemini-3-7-flash-medium-vs-grok-4-6). At the same high reasoning setting, Grok scores 61 to Gemini's 56 on its Intelligence Index. Gemini produces 323 output tokens per second to Grok's 56, while costing $0.40 per Index task to Grok's $0.84.
_Source: [Artificial Analysis](https://artificialanalysis.ai/models/comparisons/gemini-3-7-flash-medium-vs-grok-4-6), captured 19 August 2026. High reasoning on both models. External benchmark data, not Chat Thing performance._
Run the same questions through both models before changing a live agent. A polished test answer is useful, but repeated behaviour matters more: retrieval, tone, citations, Power-Up decisions and how the model handles uncertainty.
## Power-Up activity, without all the clutter
When an agent used several Power-Ups in one turn, each call could appear as its own row. Longer tasks left a trail of repeated chips, empty gaps and loading indicators that jumped around while the answer was being built.
Those calls now collapse into one summary for the whole turn. You can see what the agent is using while it works, then a tidy record of what it used when it finishes. Team owners can expand the summary to inspect every individual call when they need the detail.
The summary moves from "Using" to "Used" when the work is complete, so you can follow progress without opening the detail.
## More file types in Google Drive sync
[Google Drive sources](https://chatthing.ai/data-sources) can now import uploaded DOCX, JSON, HTML and EPUB files.
Previously, some of those files could sit in a connected Drive folder without ever being discovered, even though Chat Thing already understood how to process parts of them elsewhere. Discovery and parsing now use the same supported-format list, with a test to stop the two drifting apart again.
PDF, plain text, Markdown and CSV support remain in place, alongside Google Docs and Sheets.
## Try the new models
Open your agent, head to its advanced model settings and choose Gemini 3.7 Flash or Grok 4.6. Model selection is available on plans that include the model picker.
Start with your real evaluation questions rather than a blank chat. If you run a high-volume support agent, try Gemini 3.7 Flash first. If you are testing a harder multi-step workflow, compare Grok 4.6 as well.
Then open a conversation that uses several Power-Ups to see the new collapsed summary in action.
### **Related articles **
[](https://chatthing.ai/blog/ask-otto-and-seven-new-models) [Otto now lives inside every agent, plus seven new models
](https://chatthing.ai/blog/ask-otto-and-seven-new-models)
Otto now lives inside every agent, so you can change things by describing what you want rather than hunting through settings forms. Plus Claude Opus 5 and six more models, four new embed capabilities, and twelve retired models.
[](https://chatthing.ai/blog/kimi-k3-lands-in-chat-thing) [Kimi K3 is live in Chat Thing: what it is, and is it actually good?
](https://chatthing.ai/blog/kimi-k3-lands-in-chat-thing)
Moonshot AI's Kimi K3 launched today and it's already live in Chat Thing. Here's what Kimi K3 is, the benchmarks behind the hype, and how it performed when we ran it head-to-head against Claude Sonnet 5, GPT-5.6 and Gemini on real support-bot prompts.
[](https://chatthing.ai/blog/gpt-5-6-and-thirteen-new-models) [GPT-5.6 lands in Chat Thing, plus twelve more new models
](https://chatthing.ai/blog/gpt-5-6-and-thirteen-new-models)
GPT-5.6 (Sol, Terra and Luna) is now live in Chat Thing, alongside twelve more new models. A 1.05M-token context window, vision and power-up support across the board, and it is unusually light on message tokens for how capable it is.
- **Previous **
- [← Connect Your Agent to Any API](https://chatthing.ai/blog/use-api)