Model ● Covered in 1 story + Follow
This profile is built automatically from TLDRocket coverage.
No specifications recorded yet.
The daily briefing
TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.
TLDRocket
30 September 2026
Google’s Gemini 4 Argon took the spotlight and, more importantly, forced the market to talk about capacity, cost, and access as part of “safety.” DeepMind’s new model is built for long-horizon software work, knowledge tasks, and cyber defense, and it can produce up to 1,000,000 output tokens in a single response—up from the earlier 64K ceiling on prior Gemini models. Google also posted a practical price card for enterprise planners: $2 per 1M input tokens and $10 per 1M output tokens (with cached inputs at $0.10 per 1M). The rollout is staged, too—first via Google’s Fairwind program for trusted cyber defenders and through a phased U.S. government pre-release process—because Argon is already being used internally.
Read the full briefing →