- Synexa
- Model explainers
- Gemini 4 Neon
AI model ยท Google
Gemini Neon, Explained: Specs, Pricing, Benchmarks
Gemini Neon is Google's new Gemini 4 model for complex coding, knowledge work and science, released November 18, 2026 at $3 per million input tokens. Synexa does not serve it; use Synexa when your app also needs images, video, audio or 3D from one API key.
- 1M-token context window
- $3 in / $15 out per MTok
- Synexa: image, video, audio, 3D APIs
What is Gemini Neon?
Gemini Neon, full name Gemini 4 Neon, is a large language model from Google DeepMind, announced on November 18, 2026. Google describes it as bringing frontier-level intelligence with fast end-to-end task completion across complex software engineering, knowledge work and science workflows. It reads text, images, audio, video and PDFs and writes text.
Gemini Neon is the middle member of the Gemini 4 family. Carbon handles the hardest, long-horizon projects, Hydrogen handles fast everyday execution, and Neon sits between them for complex, high-frequency work. Google's model card says Neon is adapted from Carbon.
The launch came six days after Hydrogen, which we cover in our Hydrogen guide. Unlike Hydrogen, Neon is not on the Gemini API free tier, and in the Gemini app it is limited to Google AI Pro and Ultra subscribers.
Gemini Neon specs: context, output and model ID
Published figures from Google's model page, developer guide and model card, as of November 2026.
- Maker
- Google (Google DeepMind)
- Released
- November 18, 2026 (generally available)
- API model ID
gemini-4-neon- Input
- Text, images, audio, video and PDF
- Output
- Text
- Context window
- 1,048,576 tokens, shared by input, thinking and output
- Max output
- 65,536 tokens per synchronous response; up to 1M with long decode
- Thinking levels
- low, medium, high (default), xhigh
- Knowledge cutoff
- March 2026
- Based on
- Adapted from Gemini 4 Carbon
- Free tier
- Not available; paid tier only
How much does Gemini Neon cost?
On the paid tier of the Gemini API, the model costs $3 per million input tokens and $15 per million output tokens, flat across the whole 1M context. Thinking tokens bill as output. Standard rates as of November 2026.
Input, per MTok
- Gemini Neon
- $3.00
- Gemini 4 Hydrogen
- $1.50
- Gemini 4 Carbon
- $2.00 to Nov 30, 2026, then $4.00
Output, per MTok
- Gemini Neon
- $15.00
- Gemini 4 Hydrogen
- $7.50
- Gemini 4 Carbon
- $10.00, then $20.00
Context caching, per MTok
- Gemini Neon
- $0.15 (implicit only)
- Gemini 4 Hydrogen
- $0.075 (implicit only)
- Gemini 4 Carbon
- $0.10, then $0.20
Free tier
- Gemini Neon
- Not available
- Gemini 4 Hydrogen
- Yes
- Gemini 4 Carbon
- Not available
Example: 100K in + 10K out
- Gemini Neon
- $0.45
- Gemini 4 Hydrogen
- $0.225
- Gemini 4 Carbon
- $0.30, then $0.60
| Rate | Gemini Neon | Gemini 4 Hydrogen | Gemini 4 Carbon |
|---|---|---|---|
| Input, per MTok | $3.00 | $1.50 | $2.00 to Nov 30, 2026, then $4.00 |
| Output, per MTok | $15.00 | $7.50 | $10.00, then $20.00 |
| Context caching, per MTok | $0.15 (implicit only) | $0.075 (implicit only) | $0.10, then $0.20 |
| Free tier | Not available | Yes | Not available |
| Example: 100K in + 10K out | $0.45 | $0.225 | $0.30, then $0.60 |
The example row is our arithmetic from the list prices: 0.1M ร $3 plus 0.01M ร $15 = $0.45. Batch and Flex cost $1.50 / $7.50 and Priority $5.40 / $27 per million tokens. Grounding with Google Search includes 5,000 free requests a month, then costs $14 per 1,000. Google says Neon has no introductory price.
Gemini 4 Neon benchmarks: how does it score?
Scores from Google DeepMind's model card, against the two rivals Google chose. We have not run them ourselves.
Artificial Analysis Intelligence Index v5.0.1
- Gemini Neon
- 61
- Claude Opus 5.5
- 59
- GPT-6.2 Sol
- 57
Vals Index v2.2
- Gemini Neon
- 70.9%
- Claude Opus 5.5
- 70.1%
- GPT-6.2 Sol
- 66.1%
CursorBench 4.1
- Gemini Neon
- 65.4%
- Claude Opus 5.5
- 58.1%
- GPT-6.2 Sol
- not listed
Terminal-Bench 4.1 (cheating-adjusted)
- Gemini Neon
- 65.2%
- Claude Opus 5.5
- 65.8%
- GPT-6.2 Sol
- 61.5%
FrontierCode 2.0
- Gemini Neon
- 46.2%
- Claude Opus 5.5
- 47.3%
- GPT-6.2 Sol
- 43.8%
SWE-sweep (provider harness)
- Gemini Neon
- 39.0%
- Claude Opus 5.5
- 10.1%
- GPT-6.2 Sol
- 8.3%
AutomationBench
- Gemini Neon
- 56.3%
- Claude Opus 5.5
- 42.5%
- GPT-6.2 Sol
- 41.3%
Agents' Last Exam v1.1
- Gemini Neon
- 75.8%
- Claude Opus 5.5
- 71.6%
- GPT-6.2 Sol
- 73.0%
HLE-Diamond (reasoning)
- Gemini Neon
- 71.3%
- Claude Opus 5.5
- 67.9%
- GPT-6.2 Sol
- 64.0%
GraphWalks, 256K to 1M (long context)
- Gemini Neon
- 95.2%
- Claude Opus 5.5
- 66.8%
- GPT-6.2 Sol
- 80.1%
| Benchmark | Gemini Neon | Claude Opus 5.5 | GPT-6.2 Sol |
|---|---|---|---|
| Artificial Analysis Intelligence Index v5.0.1 | 61 | 59 | 57 |
| Vals Index v2.2 | 70.9% | 70.1% | 66.1% |
| CursorBench 4.1 | 65.4% | 58.1% | not listed |
| Terminal-Bench 4.1 (cheating-adjusted) | 65.2% | 65.8% | 61.5% |
| FrontierCode 2.0 | 46.2% | 47.3% | 43.8% |
| SWE-sweep (provider harness) | 39.0% | 10.1% | 8.3% |
| AutomationBench | 56.3% | 42.5% | 41.3% |
| Agents' Last Exam v1.1 | 75.8% | 71.6% | 73.0% |
| HLE-Diamond (reasoning) | 71.3% | 67.9% | 64.0% |
| GraphWalks, 256K to 1M (long context) | 95.2% | 66.8% | 80.1% |
Vendor-reported numbers on Google's own setups. Gemini Neon's SWE-sweep score is in the Google Antigravity harness; on the neutral mini-SWE harness it scores 12.8% against 3.6% and 11.6%. Treat all of it as a direction, then run your own evals.
What is Gemini Neon good at?
Google pitches the new model as frontier quality at a lower cost per task than the larger models.
Agentic coding
Lower cost per task
Knowledge work
Cyber defense
Disciplined agents
Orchestrator for Hydrogen
Where can you use Gemini Neon?
Google began rolling Gemini Neon out on launch day across developer, business and consumer products.
Gemini API and AI Studio
gemini-4-neon on the paid tier of the Gemini API from Google AI Studio or Android Studio. Google AI Pro and Ultra subscribers can also try it in AI Studio.Gemini for business
Gemini app
Google Antigravity
Fairwind Program
What to check before you switch
Details from Google's migration guide and pricing page that affect what Gemini Neon costs and how it behaves.
- Start at the default high level and test medium; keep xhigh for hard multi-step work
- Set background mode and max_output_tokens above 65,536 for long outputs
- Raise HTTP client timeouts to at least 15 minutes for hard prompts
- Budget for thinking tokens, which are billed at the output rate
- Plan on explicit context caching; only implicit caching is supported
- Expect a free tier; the model is paid tier only
- Send the
minimalthinking level, or rely on temperature and top_p - Expect knowledge past March 2026 without grounding or search
Can you run Gemini Neon on Synexa?
No. Synexa does not serve Gemini Neon or any other Gemini language model, and it offers no chat LLMs; its only text outputs are LLaVA-13B image answers, Whisper speech transcripts and an NSFW image-check label. For the language side, use the Gemini API or Gemini for business.
Synexa covers the media half of an AI product through one REST API. One key gives you image models such as Nano Banana Pro at $0.10 a request, music with Lyria 2 at $0.10 a request, plus video, speech, upscaling and image-to-3D models such as Hunyuan3D 3.1 at $0.16 a request. A common pattern: let the language model plan and write the prompt, then send it to Synexa to render the asset. Prices are on each model page in Explore, new accounts start with $0.10 of free credit, and the quickstart shows the first call.
- Pay per request, no subscription
- Python and Node clients, plus plain HTTP
- Image, video, audio, upscaling and 3D in one catalog
Synexa models to pair with a Gemini Neon agent
Flat per-request prices as listed on Synexa model pages, November 2026.
Product or marketing image from a written brief
- Synexa model
- Nano Banana Pro
- Price
- $0.10 / request
Chart, slide or poster with sharp, readable text
- Synexa model
- GPT Image 2
- Price
- $0.10 / request
30-second music clip from a text prompt
- Synexa model
- Lyria 2
- Price
- $0.10 / request
Textured 3D model from one image
- Synexa model
- Hunyuan3D 3.1
- Price
- $0.16 / request
Fast draft 3D mesh
- Synexa model
- TRELLIS
- Price
- $0.02 / request
Upscale or restore a photo
- Synexa model
- Real-ESRGAN
- Price
- $0.0015 / request
Transcribe a meeting or voice note
- Synexa model
- Whisper v3
- Price
- $0.004 / request
| Job your agent hands off | Synexa model | Price |
|---|---|---|
| Product or marketing image from a written brief | Nano Banana Pro | $0.10 / request |
| Chart, slide or poster with sharp, readable text | GPT Image 2 | $0.10 / request |
| 30-second music clip from a text prompt | Lyria 2 | $0.10 / request |
| Textured 3D model from one image | Hunyuan3D 3.1 | $0.16 / request |
| Fast draft 3D mesh | TRELLIS | $0.02 / request |
| Upscale or restore a photo | Real-ESRGAN | $0.0015 / request |
| Transcribe a meeting or voice note | Whisper v3 | $0.004 / request |
Synexa catalog prices, not Google's. Video models such as Veo 3.1 are also in the catalog; check each video model page for its current price.
Gemini Neon vs Carbon and Hydrogen: which should you use?
Use Neon when Hydrogen is not strong enough and you do not need Carbon's extra verification: repository-level debugging, dense legal or finance documents, and science workflows. Google's own advice is to keep Hydrogen as the default for everyday coding and agents, and Carbon for the hardest architecture, security and science problems.
The price comparison changes on December 1, 2026. Until November 30, Carbon's introductory $2 / $10 is cheaper per token than Gemini Neon's $3 / $15; from December 1, Carbon costs $4 / $20, so Neon lists at 75% of Carbon's price. Hydrogen stays at half of Neon's rate. Per-token price is only part of the bill: Artificial Analysis measured Gemini Neon at xhigh as very verbose, so test the high level first. If your product also generates images, music or 3D assets, compare media APIs in our WaveSpeed AI alternative and Runware alternative pages, and check per-request prices on pricing.
Related pages
More model explainers and API comparisons.
Frequently asked questions
What is Gemini Neon?
Gemini Neon, officially Gemini 4 Neon, is a Google DeepMind language model released November 18, 2026. It targets complex coding, knowledge work and science workflows, reads text, images, audio, video and PDFs, and holds a 1M-token context. It sits between Gemini 4 Hydrogen and Gemini 4 Carbon in the Gemini 4 family and is adapted from Carbon.
How much does Gemini Neon cost?
On the paid Gemini API tier it costs $3 per million input tokens and $15 per million output tokens, flat across the full 1M context, as of November 2026. Thinking tokens are billed as output. Cached input costs $0.15 per million, a 95% discount. Batch and Flex halve the rates to $1.50 and $7.50. A call with 100K tokens in and 10K out costs about $0.45.
Is Gemini Neon free?
No. Google lists Gemini Neon as not available on the Gemini API free tier, so you need a paid tier project to call it. In the Gemini app and Google Antigravity it is limited to Google AI Pro and Ultra subscribers, who can also try it in Google AI Studio. If you need a free Gemini 4 model, Gemini 4 Hydrogen is on the free tier.
What is the difference between Gemini Neon, Carbon and Hydrogen?
Carbon is Google's most capable reasoning model for the hardest problems, Hydrogen is the fast, cheaper everyday model, and Gemini Neon sits between them. Neon reasons deeper than Hydrogen and finishes faster than Carbon. All three share a 1M context window and the same request schema, so you can route each step of a workflow to a different model.
What model ID does the API use?
The Gemini API model ID is gemini-4-neon, available in a stable version only. It works with both generateContent and the Interactions API. Synchronous responses return up to 65,536 tokens; with background long decode a response can run up to the 1M-token window. The default thinking level is high, and xhigh is available for harder tasks.
How does Gemini 4 Neon compare with GPT-6.2 Sol?
Google's model card puts Gemini Neon ahead of GPT-6.2 Sol on most of the benchmarks it lists, for example 61 against 57 on the Artificial Analysis Intelligence Index. GPT-6.2 Sol is cheaper per token at $2 / $10 against $3 / $15, and Artificial Analysis measured a lower cost per index task for Sol. Run both on your own workload.
What is the knowledge cutoff?
Google's model card gives a knowledge cutoff of March 2026, and notes that the model's knowledge may be less up to date in some domains. For anything newer, turn on grounding with Google Search, which includes 5,000 free search requests a month shared across Gemini 3 and newer models, then costs $14 per 1,000 requests, with a daily cap of 1,500.
Can I use Gemini Neon through Synexa?
No. Synexa does not serve any Gemini model or any other chat language model. Synexa is a pay-per-request API for generative media: images, video, music and speech, upscaling and image-to-3D. Many teams call a language model for planning and writing, then call Synexa to render the image, clip, track or 3D model. New Synexa accounts get $0.10 of free credit.
Pair Gemini Neon with a media API
Use Gemini Neon for the reasoning, then generate the images, video, music and 3D models it plans on Synexa. Pay per request, start with $0.10 free credit.