Skip to content
NewA calmer, faster ZeroHour
StoryModel releaseinfo

Google ships closed Gemini 4 Neon at $3/$15 with 1M context

6 sources · 9 articlesFirst reported Updated ()3
What's new

November 19: Mei’s Models Engineering Index ranks Gemini 4 Neon first at 69.5, ahead of Fable 5.2 at 65.8 and Opus 5.5 at 63.1.

Merged summary · x-ai/grok-4.7 · rewritten as coverage arrivesSummaries by Vani AI

Closed Gemini 4 Neon is live on paid Google tiers; Mei now ranks it first on an engineering-task index.

Google released closed-weight Gemini 4 Neon on November 18, 2026, its fourth Gemini 4 model in about a month, on the paid Gemini API, AI Studio, business Gemini, Antigravity, and the Pro and Ultra app. List price is $3 per million input tokens and $15 output, with cached input at $0.15 and about a 1 million-token window; image and audio generation are unsupported. Google cites an Artificial Analysis score of 61 at xhigh and Coding Agent Index leadership at 58.6, while The Decoder says Neon still trails GPT-6.2 Astra and Claude Fable 5.2 on that index and later cites 59 in Antigravity. On November 19, Mei’s Models Engineering Index ranked Neon first at 69.5. Speed figures conflict: Hacker News reports 127 tokens per second versus Carbon’s 46 on OpenRouter; The Decoder reports about 110 versus 57.

Key facts
  • List price is $3 input and $15 output per million tokens; cached input is $0.15.
  • Context is about 1 million tokens; default output is 64K, up to 1M.
  • Weights stay closed and Google-hosted; image and audio generation are unsupported.
  • Carbon and Argon intro rates are $2/$10, rising to $4/$20 on December 1.
  • Mei ranks Neon 69.5, Fable 5.2 65.8, and Opus 5.5 63.1.
  • Latent Space says Argon was retired; Willison says the lineup image omits it.
Explain this AI news with Vani AIWhat was announced, what is new and why it matters, in a new chat

Timeline

dated events in the reporting · oldest first
  1. Nov 18, 2026Google releases Gemini 4 Neon on paid tiers · Google DeepMind
  2. Nov 18, 2026OpenRouter serves Neon at 127 tokens per second · Hacker News
  3. Nov 19, 2026Mei ranks Neon first at 69.5 · Hacker News
  4. Dec 1, 2026Carbon and Argon intro rates rise to $4/$20 · The Decoder

What each outlet adds

beyond the shared facts
  • The DecoderPromo-versus-list pricing and about 110 versus 57 tokens per second
  • Google DeepMindRollout surfaces, SWE-sweep scores, and cyber-benchmark claims
  • MarkTechPostClosed weights, March 2026 cutoff, and Gray Swan claim
  • Simon WillisonPricing-order confusion and Argon omitted from lineup image
  • Latent SpaceArgon retired; Arena and Vals place Neon second

Coverage

oldest first · one row per article · UTC
  1. · 1d ago The DecoderFirst report

    Google releases Gemini 4 Neon, fourth Gemini 4 model in a month

    Neon is priced at $3 per million input tokens and $15 per million output tokens, above Hydrogen, Carbon, and Argon.

  2. · 1d ago Hacker News · AI

    Google introduces Gemini 4 Neon for coding, knowledge work, and science

    Pricing is $3 per million input tokens and $15 per million output tokens, with a 1 million-token window.

  3. · 1d ago Google DeepMind

    Google introduces Gemini 4 Neon for coding, knowledge work, and science

    It offers a 1 million-token context window at $3 per million input tokens and $15 per million output tokens.

  4. · 1d ago Simon Willison

    Google introduces Gemini 4 Neon as llm-gemini 0.38 ships

    Willison says the stated price order is Carbon, Neon, Hydrogen, while Carbon is currently half price.

  5. · 1d ago Hacker News · AI

    Author projects Gemini 4 Neon near Opus 5.5

    The projection puts Neon at Elo 1,066 versus Opus 5.5 at 1,078, with a wide interval.

  6. · 1d ago MarkTechPost

    Google DeepMind releases Gemini 4 Neon with a 1M-token context

    Weights stay closed and Google-hosted; Google reports a 61 Intelligence Index score at $3/$15.

  7. · 1d ago Latent Space

    Google DeepMind ships Gemini 4 Neon, fourth Gemini 4 model in seven weeks

    Neon lists at $3/$15 per million tokens with a 1M context window and is already on paid Gemini tiers.

  8. · 23h ago The Decoder

    Google DeepMind releases Gemini 4 Neon, its fourth model in seven weeks

    Neon has a one-million-token window, trails GPT-6.2 Astra and Claude Fable 5.2, and undercuts Carbon on list price.

  9. · 11h ago Hacker News · AI

    Gemini 4 Neon leads the Models Engineering Index at 69.5

    Anthropic's Fable 5.2 is second at 65.8 and Opus 5.5 third at 63.1 on the engineering-task index.