Google · Available now
Gemini 4 Neon (Google Neon, 2026): price, specs, benchmarks and where to use it
Gemini 4 Neon, which people also search as Google Neon, is the Gemini 4 model Google released on November 18, 2026 for complex coding, knowledge work and science, with a 1M-token context window and a price of $3 per million input and $15 per million output tokens.
Use it now
Chat
≈ 4 credits per reply · See all models
New accounts get 5 free credits once, valid for 7 days. They work on the lower-priced models here; the priciest ones need credits you bought.
Start chatting — 5 free credits- Maker
- Type
- Text
- Released
- 2026-11-18
- OpenRouter id
- google/gemini-4-neon
Per million tokens
Price
| - | Input | Output | Source |
|---|---|---|---|
| Google — Standard | $3 | $15 | aistudio.google.com |
| Google — Batch | $1.5 | $7.5 | aistudio.google.com |
| Google — Flex | $1.5 | $7.5 | aistudio.google.com |
| Google — Priority | $5.4 | $27 | aistudio.google.com |
| Via OpenRouter (google/gemini-4-neon) | $3 | $15 | openrouter.ai |
Checked
Facts, with sources
- Google announced Gemini 4 Neon on November 18, 2026 in a post by Koray Kavukcuoglu (SVP, Google DeepMind and Chief AI Architect). It describes Neon as "frontier-level intelligence with fast end-to-end task completion" across software engineering, knowledge work and science. The model card is dated the same day.Source: blog.google · checked 2026-11-19
- The Gemini API release notes list gemini-4-neon as generally available (GA) on November 18, 2026, on the paid tier at $3.00 per million input tokens and $15.00 per million output tokens, with four thinking levels (low, medium, default high, xhigh), a unified 1M-token context window and long decode up to 1M tokens.Source: aistudio.google.com · checked 2026-11-19
- Google’s deprecations page lists gemini-4-neon with release date November 18, 2026 and "No shutdown date announced". Google Cloud’s model page gives its launch stage as GA.Source: aistudio.google.com · checked 2026-11-19
- Google’s model card says Neon is built for complex coding, analytical reasoning and production-grade enterprise workflows. It says Carbon takes on the hardest long-horizon projects and Hydrogen handles fast everyday execution, while Neon "balances depth and efficiency". The card says Neon is adapted from Gemini 4 Carbon, and lists long-running agent workflows, tool-heavy tasks, sub-agent work, cyber defence tasks, repository coding and debugging, and multimodal knowledge work as intended uses.Source: deepmind.google · checked 2026-11-19
- Model page: model code gemini-4-neon (Stable); 1,048,576-token context window shared by input and output; max output 65,536 tokens by default, up to 1,048,576 with long decode; input text, image, video, audio and PDF, output text only; thinking levels low, medium, high and xhigh; function calling, structured outputs, code execution, computer use, file search, URL context, Google Search and Google Maps grounding, and agentic video understanding. Not supported: audio or image generation and the Live API. Implicit caching only (no explicit caching). Batch API, Flex and Priority inference are supported.Source: aistudio.google.com · checked 2026-11-19
- Model card: knowledge cutoff March 2026, "though in some domains the model’s knowledge may be less up to date". Google lists hallucinations as a general limitation and says it keeps working on jailbreak resistance. The card says the full Safety & Security Report for Neon "will be available in the coming days".Source: deepmind.google · checked 2026-11-19
- Google Cloud’s model page (titled "Gemini 4 Neon | Gemini Enterprise Agent Platform", at a Vertex AI documentation address) says custom temperature, top_p and top_k values are ignored (fixed at 1.0, 0.95 and 64), that thinking_budget or thinking_level "MINIMAL" returns an HTTP 400 error, and that implicit caching needs at least 4,096 tokens. It says Neon also accepts Microsoft Office and OpenDocument files, in addition to PDF and plain text.Source: cloud.google.com · checked 2026-11-19
- Google’s pricing page lists Standard at $3.00 per million input tokens and $15.00 per million output tokens (thinking included), with cached input at $0.15 (implicit caching only); Batch and Flex at $1.50 / $7.50 (cached $0.075); Priority at $5.40 / $27.00 (cached $0.27). Neon is "not available" on the Gemini API free tier. On the paid tier Google says content is not used to improve its products, and every token the model processes is billed, thinking tokens included. The page lists no date-limited price and no separate rate above 200k tokens for Neon.Source: aistudio.google.com · checked 2026-11-19
- Google’s launch post gives the same price: $3 per million input tokens, $0.15 per million cached input tokens (which it calls a 95% discount) and $15 per million output tokens, with a 1 million-token context window.Source: blog.google · checked 2026-11-19
- Google’s post says Neon is rolling out in the Gemini API in Google AI Studio (paid tier) and Android Studio, in Gemini for business (via API and model dropdown), in Google Antigravity for paid Google AI Pro and Ultra subscribers, and in the Gemini app and AI Studio for Pro and Ultra subscribers. A cyber version is rolling out "in the coming days" to eligible members of its Fairwind Program, who request access through their Cloud account team.Source: blog.google · checked 2026-11-19
- Google’s model card lists the channels as the Gemini app, Gemini API, Google AI Studio, Gemini for business and Google Antigravity, and says access varies by channel and subscription tier. It does not name Vertex AI in that list.Source: deepmind.google · checked 2026-11-19
- Google Cloud’s model page (a Vertex AI documentation address, now titled Gemini Enterprise Agent Platform) lists gemini-4-neon as GA, released November 18, 2026, knowledge cutoff March 2026, in the global location and the us and eu multi-regions, with Standard and Priority pay-as-you-go, batch inference and Provisioned Throughput (no deferred tier). It says Neon is available to verified enterprise customers in select regions, and marks it a Google Frontier Model subject to Section 34 of the Service Specific Terms (abuse monitoring).Source: cloud.google.com · checked 2026-11-19
- Google’s own figures, not tested by us: for cyber work the post reports 51.26% on WebExploitBench v1.2, which it calls state of the art at half the cost of competing large models, and says Neon for cyber is going to trusted defenders through the Fairwind Program. The model card’s safety tables compare Neon only with Gemini 4 Hydrogen: text, multilingual and image-to-text safety within confidence intervals, and unjustified refusals 3.00 percentage points lower.Source: blog.google · checked 2026-11-19
- Listed on OpenRouter as google/gemini-4-neon (listed November 18, 2026) with a 1,048,576-token context window, audio, file, image, text and video input and text output, at $3 / $15 per million input / output tokens (checked November 19, 2026). OpenRouter shows reasoning as always on, with effort levels low, medium, high and xhigh.Source: openrouter.ai · checked 2026-11-19 · second-hand
Gemini 4 family
Gemini 4 Neon next to Hydrogen and Carbon
- Google’s what’s-new page says Neon sits between Gemini 4 Hydrogen and Gemini 4 Carbon for high-frequency complex tasks. It recommends Neon for complex software engineering, knowledge work and science where you want frontier-level intelligence with fast end-to-end task completion, Carbon for the hardest architecture, debugging, science and security problems, Hydrogen as the default for everyday coding and agentic work, and Gemini 3.5 Flash-Lite for high-volume, low-latency tasks.Source: aistudio.google.com · checked 2026-11-19
Side by side
Gemini 4 Hydrogen, Neon and Carbon compared
Neon is the middle one of the three. Google says it "balances depth and efficiency": Hydrogen is for fast everyday execution and Carbon for the hardest long-horizon work. Through November 30, 2026, Neon is the dearest per token of the three on Google’s Standard list; from December 1 Google lists Carbon higher. Google also lists Gemini 4 Argon, which this table leaves out. For the other two see Gemini 4 Hydrogen and Gemini 4 Carbon, and for the whole product line Google Gemini. A rumoured Krypton model is covered separately on Gemini 4 Krypton.
| What | Gemini 4 Hydrogen | Gemini 4 Neon | Gemini 4 Carbon |
|---|---|---|---|
| How Google describes it | Speed and efficiency for everyday coding and agentic workflows | Frontier-level intelligence with fast end-to-end task completion | Its strongest reasoning model, for sustained agentic coding, science and cyber defence |
| Generally available in the Gemini API | November 12, 2026 | November 18, 2026 | November 10, 2026 |
| Standard price per million tokens (input / output) | $1.50 / $7.50 | $3 / $15 | $2 / $10 through November 30, 2026; $4 / $20 from December 1, 2026 |
| Gemini API free tier | Yes (Google may use the content to improve its products) | Not available | Not available |
| Thinking levels | low, medium, high (default), xhigh | low, medium, high (default), xhigh | low, medium, high |
| Context window and max output | 1M tokens; output 64K, up to 1M with long decode | 1M tokens; output 64K, up to 1M with long decode | 1M tokens; output 64K, up to 1M with long decode |
| Input it reads | Text, image, audio, video, PDF | Text, image, audio, video, PDF | Text, image, audio, video, PDF |
Source: aistudio.google.comSource: aistudio.google.comSource: aistudio.google.comSource: aistudio.google.comSource: aistudio.google.comSource: aistudio.google.com
Google’s own figures
Gemini 4 Neon benchmarks: what Google prints
- Google’s own figures, partly from third-party indexes it cites, not tested by us: Artificial Analysis Intelligence Index 61 at xhigh thinking (#3 overall); Vals Index #2 at 70.9%; Artificial Analysis Coding Agent Index 58.6 (#1); Terminal-Bench 4.1 65.2%; FrontierCode 2.0 46.2% at the default high setting. At high, Google says cost per task is 32% lower than Carbon.Source: blog.google · checked 2026-11-19
- Google’s model card compares Neon with Claude Opus 5.5 and GPT-6.2 Sol on 19 benchmark rows (results as of November 2026, Google’s own table). Neon scores higher on most rows; Claude Opus 5.5 scores higher on Terminal-bench 4.1 and FrontierCode 2.0, and GPT-6.2 Sol scores higher on Chartography. The card points to Google’s evaluation-methodology page for details.Source: deepmind.google · checked 2026-11-19
Google’s own table
Gemini 4 Neon, Claude Opus 5.5 and GPT-6.2 Sol on Google’s model card
These are the maker’s own numbers, so read them as Google’s case for Neon, not as a neutral ranking. A dash means the card shows no score for that model. The rivals are Claude Opus 5.5 and GPT-6.2 Sol; the table is a selection of the card’s 19 rows, and in three rows (Terminal-bench 4.1, FrontierCode 2.0, Chartography) a rival scores higher than Neon.
| Benchmark | Gemini 4 Neon | Claude Opus 5.5 | GPT-6.2 Sol |
|---|---|---|---|
| AA Index v5.0.1 (capability index) | 61 | 59 | 57 |
| Vals Index v2.2 (capability index) | 70.9% | 70.1% | 66.1% |
| CursorBench 4.1 (agentic coding, score) | 65.4% | 58.1% | — |
| Terminal-bench 4.1 (agentic coding, cheating-adjusted accuracy) | 65.2% | 65.8% | 61.5% |
| FrontierCode 2.0 (agentic coding, main score) | 46.2% | 47.3% | 43.8% |
| SWE-sweep, provider harness (agentic coding) | 39.0% | 10.1% | 8.3% |
| AutomationBench (knowledge work) | 56.3% | 42.5% | 41.3% |
| Agents’ Last Exam v1.1 (knowledge work) | 75.8% | 71.6% | 73.0% |
| HLE-Diamond (reasoning) | 71.3% | 67.9% | 64.0% |
| Chartography (multimodal) | 73.5% | 66.3% | 75.1% |
| GraphWalks, 256k to 1M, BFS (long context, F1) | 95.2% | 66.8% | 80.1% |
| WebExploitBench v1.2 (cybersecurity) | 51.3% | 37.1% | — |
Nearest available
Models you can use now instead
Gemini 4 Hydrogen
Google$1.5 / $7.5 (per million tokens)
- Google announced Gemini 4 Hydrogen on November 12, 2026 (the post was last updated November 16, 2026), and its model card is dated the same day.Source: blog.google · checked 2026-11-17
- The Gemini API changelog lists gemini-4-hydrogen as generally available (GA) on November 12, 2026, with a new xhigh thinking level and as the default model for the Antigravity agent in Gemini Managed Agents; it points to migration guidance from Gemini 3.8 Flash.Source: aistudio.google.com · checked 2026-11-17
- Google describes Gemini 4 Hydrogen as a fast, efficient model for everyday coding and agentic work, built for "rapid execution across diverse agentic workflows". Gemini 4 Carbon and Gemini 4 Argon remain its frontier models for long-horizon reasoning.Source: blog.google · checked 2026-11-17
Gemini 4 Carbon
Google$2 / $10 (per million tokens)
- Announced on October 28, 2026 for a subset of cyber defenders, with developers and consumers to follow.Source: blog.google · checked 2026-11-11
- The Gemini API changelog lists gemini-4-carbon as generally available (GA) on November 10, 2026.Source: aistudio.google.com · checked 2026-11-11
- Google’s model card lists multi-step coding, scientific research, knowledge work and cyber defense.Source: deepmind.google · checked 2026-11-11
Gemini 4 Argon
Google$2 / $10 (per million tokens)
- Google announced Gemini 4 Argon on September 30, 2026, saying it was rolling out to trusted cyber defenders through its Fairwind Program, with release to developers, enterprises and consumers to follow "as soon as possible", starting with paid API customers and Google AI Ultra subscribers. The post gives no date and does not say how long the introductory price lasts.Source: blog.google · checked 2026-11-16
- The Gemini API changelog lists gemini-4-argon as generally available (GA) on October 13, 2026.Source: aistudio.google.com · checked 2026-11-16
- Google’s model card (published October 13, 2026) lists these channels: the Gemini app, Gemini API, Google AI Studio, Gemini for business, Google Antigravity and Google AI Mode. Google’s pricing page says Gemini 4 Argon is not available on the Free tier.Source: deepmind.google · checked 2026-11-16
FAQ
Questions
What is Gemini 4 Neon, and is it the same as Google Neon?
Google’s own pages call it Gemini 4 Neon (model code gemini-4-neon); "Google Neon" is simply how many people search for it. Google describes it as frontier-level intelligence with fast end-to-end task completion for software engineering, knowledge work and science, adapted from Gemini 4 Carbon. It reads text, image, video, audio and PDF and answers in text. You can chat with it on this page on credits, without a Google account.
How much does Gemini 4 Neon cost?
Google’s pricing page lists Standard at $3 per million input tokens and $15 per million output tokens (cached input $0.15), Batch and Flex at $1.50 / $7.50, and Priority at $5.40 / $27. OpenRouter lists google/gemini-4-neon at $3 / $15 (checked November 19, 2026). On this site a typical reply to this model costs about 4 credits (a credit is about $0.01 in the $5 pack), billed by its real token cost, at least 1 credit.
How do I use Gemini 4 Neon?
Google lists the Gemini API in Google AI Studio (paid tier, model code gemini-4-neon) and Android Studio, Gemini for business, Google Antigravity for paid AI Pro and Ultra plans, and the Gemini app for Pro and Ultra subscribers. Google Cloud’s model page says it is available to verified enterprise customers in select regions. You can also reach it through OpenRouter as google/gemini-4-neon, or try it in the chat box on this page.
How is Gemini 4 Neon different from Hydrogen, Carbon and Argon?
Google says Neon sits between Hydrogen and Carbon: Hydrogen for fast everyday work, Carbon for the hardest long-horizon problems, Neon for complex, high-frequency tasks. Neon costs more per token than Hydrogen, and more than Carbon’s $2 / $10 price today (Google lists Carbon at $4 / $20 from December 1). Google’s speed and cost claims are its own; we did not test them. See Gemini 4 Hydrogen, Gemini 4 Carbon and Gemini 4 Argon. A rumoured Krypton model is covered separately on Gemini 4 Krypton; it is not on Google’s list of Gemini 4 models.
What is the Gemini 4 Neon context window and knowledge cutoff?
Google’s model page lists 1,048,576 tokens shared by input and output. Max output is 65,536 tokens by default, or up to 1,048,576 with long decode. The price is the same across the whole window. Google’s model card gives a March 2026 knowledge cutoff, with the warning that some domains may be less up to date.
Is there a free way to use Gemini 4 Neon?
Google’s pricing page lists Neon as not available on the Gemini API free tier, so API use is on the paid tier. In the Gemini app and Google AI Studio, Google says it is for Pro and Ultra subscribers. On this site you pay in credits for each reply. Plans and limits can change, so check Google’s pages.
How good is Gemini 4 Neon?
By Google’s own figures it scores 61 on the Artificial Analysis Intelligence Index at xhigh thinking and 65.2% on Terminal-bench 4.1. Google’s model card compares it with Claude Opus 5.5 and GPT-6.2 Sol: Neon leads on most rows, but Opus 5.5 is higher on Terminal-bench 4.1 and FrontierCode 2.0. We did not test any of it; see the table on this page. Compare Claude Opus 5.5 and GPT-6.2.
Is this the official Google website?
No. This page is unofficial and not affiliated with or endorsed by Google. Facts link to the pages we read.