Google launches AI model Gemini 4 Neon for coding and research

The model is available to AI Pro and AI Ultra subscribers in the Gemini app. In independent testing, Neon finished third overall, one point ahead of Gemini 4 Carbon at the settings tested.

Numbers in brackets link to the sources listed below.

Google began rolling out Gemini 4 Neon on 18 November for software engineering, knowledge work and science[1GoogleIntroducing Gemini 4 NeonOpen sourceSources [1]]. AI Pro and AI Ultra subscribers can use the new AI model in the Gemini app, where it replaces Argon[29to5GoogleGoogle launches Gemini 4 Neon with Carbon quality at Flash speeds, live for AI ProOpen sourceSources [2]].

ArtificialWatch issued a launch alert after verifying that Neon was live on Google’s Vertex AI platform[5ArtificialWatchGemini 4 Neon is live — news & launch alertsOpen sourceSources [5]]. The outlet later confirmed that developers had general access to Neon through the paid tier of the Gemini API[5ArtificialWatchGemini 4 Neon is live — news & launch alertsOpen sourceSources [5]].

Google DeepMind says it adapted Neon from Gemini 4 Carbon and that the model accepts text, images, audio and video[3Google DeepMindGemini 4 Neon - Model CardOpen sourceSources [3]]. The input and the model’s reply share a limit of one million tokens[3Google DeepMindGemini 4 Neon - Model CardOpen sourceSources [3]]. Tokens are the units a model uses to represent and process information. Google charges $3 for a million input tokens and $15 for the same volume of output, with cached input priced at $0.15 per million[1GoogleIntroducing Gemini 4 NeonOpen sourceSources [1]].

Artificial Analysis ranked Neon third on its Intelligence Index with a score of 61 at the xhigh reasoning setting, one point above Carbon at high[4Artificial AnalysisGoogle continues to accelerate its release cadence with Gemini 4 NeonOpen sourceSources [4]]. At those settings, Neon took about 21 minutes per task compared with about 36 minutes for Carbon, a difference the testers attributed to Neon’s faster output[4Artificial AnalysisGoogle continues to accelerate its release cadence with Gemini 4 NeonOpen sourceSources [4]].

Sources

  1. 1
    Introducing Gemini 4 Neon

    PrimaryGoogleAccessed

  2. 2
  3. 3
    Gemini 4 Neon - Model Card

    PrimaryGoogle DeepMindAccessed

  4. 4
    Google continues to accelerate its release cadence with Gemini 4 Neon

    IndependentArtificial AnalysisAccessed

  5. 5
    Gemini 4 Neon is live — news & launch alerts

    IndependentArtificialWatchAccessed

ITANU’s newsroom system produced this article with AI and checked it against the sources listed above. How we work

ITANU Magazine

AI, computing and science news