ReleasesOct 1, 2026 · 03:228 sources
Google Launches Gemini 4 Argon with 1M Token Output Limit
From huggingnews.com
Google has launched Gemini 4 Argon, a frontier AI model specialized in software engineering and cyber defense, with a 1M token output limit and initial rollout to cybersecurity partners and internal teams via the Fairwind Program. The model scored 77.9% on the DeepSWE v1.1 benchmark, offers $2 per 1M input tokens and $10 per 1M output tokens, and is being coordinated with the US government before expanding to paid API customers and AI Ultra subscribers. Internal deployments have optimized a quantum computing subroutine by 40% and freed 300TiB to 1PiB of memory across Google's data centers.
Read the sourceSign in to suggest
Sources
+ 6 more
- SUPPORT@firstsquawk“release Argon without cyber guardrails, and that it is deploying misalignment mitigations”x.com
- SUPPORT@wallstengine“generate up to 1 million output tokens in a single trajectory, up from 64K previously”x.com
- SUPPORT@valsai“in the top 5 on 20 of the 22 benchmarks we ran”x.com
- SUPPORT@firstadopter“Google's models have historically tended to be bench-maxxed, performing better on benchmarks than in real world use”x.com
- SUPPORT@mweinbach“used it to profile and optimize their data centers globally, saving 300TB of memory and up to 1PB”x.com
- SUPPORT@spicey_lemonade“along with Fable 5.1, GPT-6 Astra, and Opus 5.5, it belongs to a small group of models capable of beating the game”x.com
#google#llmMarkdown
Content history (1)
- 2026-10-04 · Summary · vi · Replaced stored textPhiên bản Argon của Gemini 4 cho phép tạo ra tối đa 1 triệu token mỗi lần gọi, mở rộng kh…Mô hình frontier AI mới của công ty hiện đang được rollout cho một nhóm đối tác an ninh m…
One ranked edition, every morning
The stories that matter today. Free, cancel anytime.
Telegram @aidr_today