İçeriğe geç / Skip to content / Zum Inhalt

Is Gemini 4 Neon Free? Pricing and Access Guide

Ahmet Balaman

10 min read

Vibe CodingGemini 4Gemini 4 NeonGemini APIGoogle AntigravityPricingAI
Is Gemini 4 Neon Free? Pricing and Access Guide

No, Gemini 4 Neon is not free. It has no free API tier and is not included in the free Gemini app plan or the free Google Antigravity tier. Using Neon requires a paid Gemini API account, at least a Google AI Pro or Google AI Ultra subscription, or an eligible enterprise plan. Gemini 4 Hydrogen remains the only Gemini 4 model with free access.

This post is a news roundup and compilation; I have not tested the model myself. The details come from Google's November 18, 2026 announcements, Google AI Studio pricing and model pages, Google Antigravity docs, and Artificial Analysis. I checked the pricing and docs pages on November 19, 2026; they can change, so verify against your own account before making decisions. Sources are at the end. For the technical overview, see What Is Gemini 4 Neon?; for benchmarks, see Neon vs Carbon vs Opus 5.5 vs GPT-6.2 Sol; and for Antigravity workflows, see What Is Antigravity Orchestrator?.

Where is it available, and to whom?

Surface Who gets access Notes
Gemini API Paid tier only (gemini-4-neon) No free tier; supports generateContent and the Interactions API
Google AI Studio Paid API developers; Google AI Pro and Ultra subscribers Google notes Pro and Ultra users can try Neon in AI Studio
Gemini app Google AI Pro and Ultra subscribers Available in the model drop-down
Google Antigravity (App, CLI, IDE) Paid Google AI Pro (or eligible device trial), Ultra, Enterprise Default model for new chats; context limit listed as 512K tokens
Antigravity Orchestrator (Neon) Paid Pro (or eligible device trial) and Ultra Neon plans, Hydrogen codes; Enterprise shown closed in docs table
Android Studio Gemini API paid tier Listed in Google's launch announcement
Gemini for business Eligible enterprise accounts Available via API, model drop-down, and Antigravity
Google Search (AI Mode) Not mentioned Not listed in the Neon launch post (Search AI Mode uses Hydrogen)
Fairwind Program Trusted cyber defenders Gemini 4 Neon for cyber rolling out over the coming days

Two details worth noting:

  • New default in Antigravity. On paid Pro, eligible device trials, Ultra, and Enterprise accounts, Neon now sits at the top of Antigravity's model selector and serves as the default model for new conversations when you haven't picked another model. Antigravity's docs list its context window there as up to 512K tokens (compared with 1,048,576 tokens in the API).
  • Managed Agents. According to AI Studio's "Choosing a model" guide, the Antigravity agent in Gemini Managed Agents still defaults to gemini-4-hydrogen, though you can override model to gemini-4-neon in agent_config.

API pricing: Standard, Batch, Flex, Priority

The gemini-4-neon rates on Google AI Studio's pricing page (USD per 1 million tokens):

Tier Input Output (including thinking) Cached input (implicit)
Standard $3.00 $15.00 $0.15
Batch $1.50 $7.50 $0.075
Flex $1.50 $7.50 $0.075
Priority $5.40 $27.00 $0.27
Free tier Not available Not available Not available

Four rules complete the table:

  • Flat rate across the full 1M window. There is no long-context surcharge and no introductory promo period for Neon.
  • Every processed token is billed. Thinking tokens, plus internal system instructions and function declarations used during grounding, count toward billed tokens.
  • Implicit caching only. Explicit caching (cached_content) is not supported; repeated prompt prefixes of at least 4,096 tokens automatically get a 95% discount via implicit caching.
  • Privacy and grounding. According to the pricing page, paid-tier prompts and outputs are not used to improve Google products. Grounding with Google Search or Google Maps includes 5,000 free requests per month (shared across Gemini 3 and newer models; Search has a 1,500 requests/day cap), then $14 per 1,000 queries.

Why the Nov 30 / Dec 1 pricing shift matters

Neon lists at $3.00 / $15.00 per million tokens. Meanwhile, Carbon and Argon are on introductory pricing of $2.00 / $10.00 through November 30, 2026, after which they double to $4.00 / $20.00 on December 1, 2026.

That creates a brief window in late November where Neon's per-token list price looks 50% higher than Carbon and Argon's promo rate. Two facts flip that picture: first, starting December 1 Neon will be 25% cheaper per token than Carbon and Argon; second, according to Artificial Analysis measurements at the default high setting, Neon uses fewer tokens per task and already costs less per task than Carbon today.

Worked cost example

Assume a single API call with 100,000 input tokens (0.1M) and 5,000 output tokens (0.005M, including thinking tokens):

Scenario Calculation Cost
Neon, Standard 0.1 × $3.00 + 0.005 × $15.00 = $0.30 + $0.075 $0.375
Neon, Batch / Flex 0.1 × $1.50 + 0.005 × $7.50 = $0.15 + $0.0375 $0.1875
Neon, Priority 0.1 × $5.40 + 0.005 × $27.00 = $0.54 + $0.135 $0.675
Neon, Standard (100% cached input) 0.1 × $0.15 + 0.005 × $15.00 = $0.015 + $0.075 $0.090
Hydrogen, Standard 0.1 × $1.50 + 0.005 × $7.50 = $0.15 + $0.0375 $0.1875
Carbon / Argon (through Nov 30) 0.1 × $2.00 + 0.005 × $10.00 = $0.20 + $0.05 $0.250
Carbon / Argon (from Dec 1) 0.1 × $4.00 + 0.005 × $20.00 = $0.40 + $0.10 $0.500

At 1,000 requests per day, Neon Standard runs $375, Neon Batch or Flex runs $187.50, and post-December 1 Carbon or Argon runs $500. Keep in mind that Neon defaults to high thinking, so complex tasks can easily produce far more than 5,000 output tokens.

How thinking levels and Nov 18 file slicing change your bill

Neon supports four thinking levels (low, medium, high, xhigh). Because thinking tokens bill at the $15/M output rate, your setting drives cost. Artificial Analysis Intelligence Index v5.0.1 figures (checked November 19, 2026):

Thinking level Index Cost per index task Time per task Output tokens
low 44 $0.75 3.7 min 28.3M
medium 55 $1.90 8.3 min 62.9M
high (default) 60 $3.74 14.8 min 123M
xhigh 61 $6.44 21.5 min 283M

Moving from medium to the default high nearly doubles task cost ($1.90 to $3.74) for +5 points. Moving from high to xhigh raises task cost by 72% ($3.74 to $6.44) for just +1 point. Benchmarking your own workload across all four levels often yields big savings.

Input savings from the Nov 18 file-slicing update: Alongside Neon, Google updated the Gemini API so you can slice PDFs, Microsoft Office and OpenDocument files, videos, and audio—passing only specific page ranges, worksheet tabs, cell ranges, time intervals, or content layers instead of paying to ingest an entire file.

Which plan gets which Gemini 4 model?

Here is the Google Antigravity plan table as of November 19, 2026 (note that the Hydrogen row also changed compared with earlier docs):

Plan Neon Carbon Argon Hydrogen Orchestrator
Free and Google AI Plus No No No No in Antigravity (free API tier & AI Mode exist) No
Google AI Pro (eligible device trial) Yes No Yes No (requires paid Pro) Neon and Argon
Google AI Pro (paid) Yes No Yes Yes Neon and Argon
Google AI Ultra Yes Yes Yes Yes Neon, Carbon, Argon
Enterprise Yes Yes Yes Yes Closed in docs table

Two points to keep in mind:

  • Hydrogen's Antigravity row changed. On November 14, the Antigravity docs table showed Hydrogen on Free and Plus plans; as of November 19, the table restricts Hydrogen in Antigravity to non-trial paid Pro, Ultra, and Enterprise, leaving Free and Plus users on Gemini 3.8 Flash.
  • Quotas and overages. Pro receives a five-hour refreshed quota up to a weekly limit; Ultra receives the highest five-hour refreshed quota. Both tiers can purchase AI credits for overages.

Watch out for fake download pages

Gemini 4 Neon is not downloadable software and has no public weights. Close any non-Google site offering a "Gemini 4 Neon installer," "free API key," or "early access login." Official access runs solely through Google AI Studio, the Gemini app, Google Antigravity, Android Studio, and Gemini for business.

Frequently Asked Questions

Can I use Gemini 4 Neon for free?

No. There is no free API tier, and both the Gemini app and Google Antigravity require at least a Google AI Pro or Ultra subscription. If you want free Gemini 4 access, use Gemini 4 Hydrogen.

How much does the Gemini 4 Neon API cost?

On the Standard paid tier it costs $3.00 per million input tokens, $15.00 per million output tokens (including thinking tokens), and $0.15 per million cached input tokens. Batch and Flex cost half ($1.50 / $7.50 / $0.075), and Priority costs $5.40 / $27.00 / $0.27.

Does Google AI Pro include Gemini 4 Neon?

Yes. Paid Google AI Pro plans (and eligible premium device trials in Antigravity) include Neon in the Gemini app and Antigravity, along with Orchestrator (Neon).

Is Neon cheaper or more expensive than Carbon?

Through November 30, 2026, Carbon's $2 / $10 introductory price is lower per token than Neon's $3 / $15 rate; starting December 1, Carbon rises to $4 / $20, making Neon 25% cheaper per token. On Artificial Analysis, Neon high ($3.74/task) is already cheaper per task than Carbon high ($4.90/task).

Are my paid API prompts used to train Google models?

According to Google AI Studio's pricing page, data on the paid tier is not used to improve Google products.

Can I try Neon for free in Google AI Studio?

API billing is paid-only, though Google notes that Google AI Pro and Ultra subscribers can try Neon inside Google AI Studio.

Sources

  • Google, Introducing Gemini 4 Neon (November 18, 2026): Pro/Ultra access across AI Studio, Gemini app, Antigravity, and Android Studio; Fairwind rollout.
  • Google AI Studio, Gemini API pricing (checked November 19, 2026): Standard, Batch, Flex, and Priority pricing, no free tier, data policy, Search/Maps grounding fees.
  • Google AI Studio, Choosing a model and release notes (checked November 19, 2026): Carbon promo end date, Managed Agents default, November 18 file slicing feature.
  • Google Antigravity, Choose a model and Plans (checked November 19, 2026): plan table, 512K context window, default model status, quota rules, and AI credits.
  • Artificial Analysis, Gemini 4 Neon and model leaderboard (checked November 19, 2026): cost, time, and token counts by thinking level.

Comments