Google DeepMind releases Gemini 4 Neon with a 1M-token context
Original title: Google DeepMind Releases Gemini 4 Neon: 1M Context, #1 on AA Coding Agent Index
Weights stay closed and Google-hosted; Google reports a 61 Intelligence Index score at $3/$15.
Neon sits between Gemini 4 Carbon and Hydrogen and is adapted from Carbon, with no disclosed parameter count. Google says it leads the AA Coding Agent Index at 58.6 and scores 95.2% on GraphWalks, while trailing Claude Opus 5.5 on FrontierCode 2.0. Paid access is through the Gemini API, Gemini for business, Antigravity, and the Gemini app for Pro and Ultra users. The launch comparison figures are Google’s own.
Unconfirmed: The launch comparison scores are reported by Google and were not independently reproduced in this article.
- Default output is 64K tokens, with up to 1M via background long decode.
- Knowledge cutoff is March 2026; inputs cover text, images, audio and video.
- Cached input is $0.15 per million tokens.
- Google reports a 0.09% misaligned-action rate and 0.0% Gray Swan prompt-injection success.
- On Vals Index and Agent Arena, the article ranks Neon second at 70.9% and 1336.
- Vendors
- Google DeepMind
From the original article
8 min read at marktechpost.comGoogle released Gemini 4 Neon, a 1M-context Gemini 4 model priced at $3/$15 per million tokens. It scores 61 on the AA Intelligence Index and ranks #1 on AA’s Coding Agent Index. The post Google DeepMind Releases Gemini 4 Neon: 1M Context, #1 on AA Coding Agent Index appeared first on MarkTechPost.
By Asif Razzaq · marktechpost.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.