Neon Fieldnotes

Multimodal reasoning / Gemini 4

Google Gemini 4 Neon AI.

Gemini 4 Neon is Google’s AI model built for fast reasoning across coding, analysis, and complex tasks. It can work with text, images, and other media.

Chat with Neon

Gemini 4 Neon

Pro and Power

Availability

Where to use Neon

Access and usage limits depend on your account and the product you open.

On this website

Explore Gemini 4 Neon AI online at gemini4neonai.com. Ask questions, review code, or explore a problem in the online Playground.

Open Playground

Gemini app

For Google AI Pro and Ultra subscribers. Sign in to the account with your plan and check the model picker.

AI Studio & API

Eligible paid plan subscribers can select Neon in AI Studio. The Developer API requires a paid project and uses the model ID gemini-4-neon.

Google API pricing

Google Antigravity

Available to eligible Pro and Ultra users. Choose Neon for coding work or check the Orchestrator option for multi-model workflows.

Gemini for business

Google announced access through the API and model dropdown. Check your organization’s license and deployment surface.

Model selection

Compare the Gemini 4 models

Start with the same set of tasks and compare completion quality, time, and total tokens.

Swipe or scroll to see all columns.

ModelStart here forExample workload
Gemini 4 HydrogenHigh-volume everyday tasksExtract structured fields from support tickets.
Gemini 4 NeonComplex coding and multimodal reasoningCompare a design brief, screenshots and a patch.
Gemini 4 CarbonDeep, long-horizon reasoningReview architecture trade-offs before a migration.
Gemini 4 ArgonAdvanced coding and deeper researchInvestigate a large repository and outline a multi-step implementation.

Gemini 4 Neon vs Argon

Choose Neon for repeated complex tasks where speed matters. If you are already using Gemini 4 Argon, compare both on your own repository or research tasks before switching.

Read the Neon vs Argon comparison

Gemini 4 Neon vs Carbon

Neon is adapted from Gemini 4 Carbon. Choose Neon for fast iteration and test Gemini Carbon on deep, long-horizon work.

Read the Neon vs Carbon comparison

How Neon scores

Artificial Analysis measured 61 on its Intelligence Index at xhigh and about 110 output tokens per second. Its model page reports 56.51 seconds to the first answer token in that setting. Output throughput and time to the first answer measure different parts of the wait.

Google’s model card reports 65.2% on Terminal-Bench 4.1, versus 65.8% for Claude Opus 5.5 and 61.5% for GPT-6.2 Sol. Check the evaluation settings when comparing scores.

Artificial Analysis · Google DeepMind model card

Common questions

Gemini 4 Neon FAQ

What is Google Neon?

People searching for Google Gemini Neon or Google Neon in the context of Gemini 4 are looking for Google Gemini 4 Neon, the multimodal reasoning model Google announced on November 18, 2026.

Gemini 4 Argon vs Neon: which should I use?

Start with Neon for frequently repeated coding or analysis tasks. Keep Argon as a baseline for complex research or existing coding workflows and compare pass rates, time and billed tokens.

Gemini Neon vs Carbon: which is faster?

Neon is positioned for faster iteration. Meaningful comparisons depend on thinking level, time to first answer, total output and the tool harness used.

Can I use Gemini 4 Neon for free?

New accounts on this site start with 25 free credits. Open the Neon Playground and sign in to use them. Google lists Neon for Pro and Ultra in its consumer app and on the paid tier of its Developer API.

What inputs and outputs does Neon support?

Google lists text, images, video, audio and PDFs as inputs, and text as output. The unified context is 1,048,576 tokens. Default synchronous output is up to 65,536 tokens, with longer output available through Long Decode subject to the context budget. Model documentation

Where can I find Gemini 4 Neon API pricing?

See the Google Gemini 4 Neon API pricing and token calculator for input, cached input, and output rates.

Why can’t I see Neon in the model picker?

Check the signed-in account and subscription, then confirm which product you’re using. For a work or school account, check with your administrator. Google announced a rollout, so individual availability may vary.

Need more room to build?

This build reached the free output limit. Try a smaller request or choose a plan with a higher output limit for larger builds.

View plans