Varun Mohan of Google DeepMind (X bio: building @antigravity @GoogleDeepMind) announced on X that the context window for the Argon, Carbon and Hydrogen models in Antigravity has been doubled, and that context compaction per user turn should drop by 3.5x. The post drew more than 5,400 likes and was reposted by Sundar Pichai and Demis Hassabis. The exact window sizes were not disclosed.
Key Takeaways
- ✓The context window for the Argon, Carbon and Hydrogen models in Antigravity has doubled, as announced on X by Varun Mohan of Google DeepMind himself
- ✓The stated effect: context compaction per user turn drops by 3.5x
- ✓The new token limits were not disclosed, and the post does not mention any pricing or quota change
- ✓The post drew more than 5,400 likes and about 330,000 views, and was reposted by Sundar Pichai and Demis Hassabis
- ✓Compaction is a common way long agentic coding sessions lose detail, so a bigger window lets the model keep more files and history in long sessions
Key Decision Metrics at a Glance
Turn your technical choice into a development budget
Compare 40 dev plans & simulate token costs vs $20/mo subscriptions
Project Links & Resources
Direct AccessIn-Depth Technical Analysis
Background
In agentic coding tools, long sessions fill up the context window. Tools usually fall back on compaction, summarizing earlier conversation and file content into a shorter digest so work can continue, but details get lost and the agent may forget earlier decisions or edits. Antigravity is Google's agentic coding platform, where Gemini 4 models such as Argon and Carbon are available.
What changed
On November 16, Varun Mohan of Google DeepMind (his X bio reads: building @antigravity @GoogleDeepMind) posted that the context window for the Argon, Carbon and Hydrogen models in Antigravity has been doubled. He said it took a lot of work and that users should see compaction per user turn drop by 3.5x. The post does not give the new token limits or say whether pricing or quotas change; these were not disclosed.
Numbers that matter
No new benchmark was released. The only figure is the 3.5x drop in compaction, stated by Varun Mohan himself. The post drew more than 5,400 likes and about 330,000 views, and was reposted by Sundar Pichai and Demis Hassabis.
Getting started
If you run long tasks, large refactors or multi-file changes in Antigravity, watch whether compaction notices become rarer and whether the agent forgets less. Keep key constraints in a project instructions file rather than relying only on long context, and check the official docs for the new window limits.
Benchmark side-by-side against alternatives, or calculate monthly token cost vs subscription break-even.
Discussion & Comments
0Sign in to join the discussion
Connect with AI developers to exchange benchmark insights.