Gemini 4 Neon rolls out with a 61 benchmark score
Gemini 4 Neon is rolling out now. Neon ranked third on the Artificial Analysis Intelligence Index, scoring 61 at its xhigh thinking level.
Key takeaways · 4
- 01
Neon placed third on the Artificial Analysis Intelligence Index with a score of 61 at xhigh thinking.
- 02
The listed Gemini app access is for Google AI Pro and Ultra subscribers; eligible Fairwind Program members were listed for access soon.
- 03
Listed API pricing is $3 per million input tokens and $15 per million output tokens.
- 04
Neon supports Computer Use, code execution, file search and grounding through Google Search and Maps, but not image or audio generation.
Rollout and access
The Gemini API in Google AI Studio, Android Studio and Gemini for business were mentioned.[1] Gemini app access was listed for Google AI Pro and Google AI Ultra subscribers, while Antigravity was listed for paid subscribers.[1] Eligible Fairwind Program members were listed for access soon.[1] These details give teams several possible routes to assess, but the evidence does not specify a single rollout schedule for every surface.
Capabilities and limits
Neon brings frontier intelligence and fast task completion to complex software engineering, knowledge work and science.[1] It handles text, images, video, audio and PDFs, and supports Computer Use, code execution, file search, and grounding through Google Search and Maps.[2] Its context window is about one million tokens, and it can produce up to one million output tokens.[2] Image generation and audio generation are not supported.[2] Teams assessing it for a workflow should distinguish its ability to process image and audio inputs from its lack of image and audio generation.
Pricing and model comparisons
The listed price for Neon is $3 per million input tokens and $15 per million output tokens.[2] Those rates are twice Hydrogen’s listed prices of $1.50 per million input tokens and $7.50 per million output tokens.[2] Carbon and Argon had an introductory rate of $2 per million input tokens and $10 per million output tokens, with rates set to double to $4 and $20 on December 1.[2] The comparisons give buyers a basis for estimating usage costs, but they do not establish which model will be less expensive for a particular workload.
What the ranking measures
Artificial Analysis says its Intelligence Index combines nine challenging evaluations spanning knowledge work, coding, science and general knowledge.[3] The organisation says it conducts all evaluations independently.[3] Neon’s third-place ranking and score of 61 are specifically reported at its xhigh thinking level.[1] The result offers a benchmark signal, rather than a guarantee of performance on any individual task; teams should compare the evaluation’s scope with their own requirements before using the ranking to make a deployment decision.
For teams evaluating models for coding, research or knowledge work, Neon combines a high-capacity context window with multiple tool capabilities and a clearly stated token price. The ranking can inform an evaluation, but the benchmark result and listed capabilities do not by themselves establish fit or cost for a specific workflow.
Why it matters
AI training for your team
kju.ai is built for teams: daily sessions, team analytics, and rollout support for your organization.
How this developed
18 November 2026
Gemini 4 Neon rolls out with a 61 benchmark score
Sources
- Google DeepMind on X: \"Neon ranks #3 with a score of 61 on the @ArtificialAnlys Intelligence Index at its xhigh thinking level. Available starting today: 🔵 The Gemini API in @GoogleAIStudio, @AndroidStudio, and Gemini for business 🔵 @Geminiapp for Google AI Pro and Ultra subscribers, and @Antigr… / Xx.com
- Google ships Gemini 4 Neon, its fourth new Gemini model in a monththe-decoder.com
- Artificial Analysis Intelligence Index v5.0.1 | Artificial Analysisartificialanalysis.ai
