All Gemini models available in 2026: Gemini 4 Carbon, Argon and Hydrogen
Updated: 2 days ago

Updated: November 17, 2026
Google's current Gemini lineup is led by three Gemini 4 models: Gemini 4 Carbon, the strongest reasoning model, Gemini 4 Argon, the efficient frontier model, and Gemini 4 Hydrogen, the fast model for everyday coding and agent work. Below them sit the Gemini 3 generation workhorses (Gemini 3.8 Flash, 3.6 Flash, 3.5 Flash-Lite and 3.1 Flash-Lite), the 3.1 Pro preview, and a set of specialized models for live voice, speech, images, video, embeddings and research agents. All three Gemini 4 models accept text, images, audio, video and PDF files and work with a context window of about 1 million tokens in the Gemini API.
Which of these you can actually pick depends on where you use Gemini. In the Gemini app, free users get Flash-Lite or automatic model selection, Google AI Plus adds Hydrogen and the Pro model, Google AI Pro adds Argon, and only Google AI Ultra lists Carbon in the model menu. In the Gemini API, Hydrogen and the Flash models have a free tier, while Argon and Carbon are paid-tier only and their introductory price of $2 input and $10 output per million tokens doubles on December 1, 2026. The tables below list every model by platform, with model codes, status and prices checked on Google's official pages on November 17, 2026, followed by a cost example for a typical app.
........
Model | Role | Gemini app | Gemini API code | API price per 1M tokens (input / output) |
Gemini 4 Carbon | Strongest reasoning model | AI Ultra | gemini-4-carbon | $2 / $10 to Nov 30, then $4 / $20 |
Gemini 4 Argon | Efficient frontier model | AI Pro, AI Ultra | gemini-4-argon | $2 / $10 to Nov 30, then $4 / $20 |
Gemini 4 Hydrogen | Fast everyday and agent model | AI Plus, AI Pro, AI Ultra | gemini-4-hydrogen | $1.50 / $7.50 |
Gemini Pro (3.1 Pro) | Advanced reasoning | AI Plus, AI Pro | gemini-3.1-pro-preview | $2 / $12 up to 200K tokens |
Gemini 3.8 Flash | Flash model for agents and enterprise work | Not in the app menu | gemini-3.8-flash | $0.75 / $3.75 to Dec 31, then $1.50 / $7.50 |
Flash-Lite (app) / Gemini 3.5 Flash-Lite (API) | Quick everyday answers | All users (version not stated) | gemini-3.5-flash-lite | $0.30 / $2.50 |
........
··········
GEMINI 4 CARBON, ARGON AND HYDROGEN: THE CURRENT FLAGSHIP MODELS.
Gemini 4 Argon was the first Gemini 4 model to reach general availability in the Gemini API, on October 13, 2026, the same day Google opened it to Google AI Pro and Ultra subscribers in the Gemini app. Google's developer documentation describes it as its most capable model for complex coding, agentic, math and science tasks at launch. Input, thinking and output share one window of 1,048,576 tokens, a single response can go past 64K tokens up to the window minus the input, and developers can choose low, medium or high thinking, with medium as the default. Argon is not available on the free tier of the API.
Gemini 4 Carbon followed as the top model of the family. Google announced on November 10, 2026 that Carbon is rolling out to all Google AI Ultra subscribers in the Gemini app, in AI Mode in Search and in Google Antigravity, and that it is available to all paying developers and enterprises through Google AI Studio and Gemini for business. A separate version for security work, Carbon for cyber, is offered only through Google's Fairwind Program for cybersecurity defenders, and scientists can apply for six months of free access to Argon and Carbon through Antigravity's highest usage tier. Google's model list describes Carbon as its strongest reasoning model for complex engineering, knowledge work, science and cybersecurity.
Gemini 4 Hydrogen is the newest of the three, launched on November 12, 2026 and rolling out globally. Google positions it as the fast, efficient model for everyday coding and agentic workflows and as an upgrade path for teams using Gemini 3.8 Flash. It is in the Gemini API and AI Studio, Android Studio, the Gemini app model menu for subscribers, Google Search for all users and Gemini in Google Sheets for AI Pro and Ultra subscribers. Free Gemini app users can receive Hydrogen answers through automatic model selection when Google's system decides it fits the prompt. Its API price of $1.50 per million input tokens and $7.50 per million output tokens is flat across the whole 1-million-token window.
··········
GEMINI MODELS IN THE GEMINI APP BY GOOGLE AI PLAN.
Google's help center lists five model names in the Gemini app: Gemini Flash-Lite for quick everyday answers, Gemini 4 Hydrogen as the balanced model for chat and tasks, Gemini Pro for advanced reasoning and coding, Gemini 4 Argon for complex workflows and Gemini 4 Carbon as the strongest model for long-horizon problems. Google's subscriptions FAQ mentions 3.1 Pro among the most capable models in the paid plans, and the help center does not give a version number for Flash-Lite. The table shows what each plan can select, as published by Google.
........
Plan (US price) | Models in chats | Models in Tasks | Context window |
No plan ($0) | Auto or Flash-Lite | Not available | 32K tokens |
Google AI Plus ($4.99 a month) | Flash-Lite, Hydrogen, Pro | Not available | 128K tokens |
Google AI Pro ($19.99 a month) | Flash-Lite, Hydrogen, Pro, Argon | Hydrogen, Argon | 1 million tokens |
Google AI Ultra (from $99.99 a month) | Flash-Lite, Hydrogen, Argon, Carbon | Hydrogen, Argon, Carbon | 1 million tokens |
........
Every plan can choose low, medium or high thinking, and the default thinking level is low. Usage is shared across models, and Google says that more advanced models and higher thinking levels use more of your allowance. Limits refresh every 5 hours until a weekly limit is reached; AI Plus has twice the standard limits, AI Pro four times, and AI Ultra 5 or 20 times the AI Pro limits depending on whether you pay $99.99 or $199.99 a month. Subscribers who run out continue the chat on Flash-Lite, and turning automatic model selection off sends every answer to Flash-Lite.
Google's subscriptions page also carries the lines "Includes Gemini 4 Hydrogen" on the free plan, "Includes Gemini 4 Argon" on AI Plus and "Includes Gemini 4 Carbon" on AI Pro, without saying in which product or with what limits. The help center's model table, which describes the Gemini app menu, does not give those plans those models, so the table above follows the help center. Work and school accounts depend on the organization's Google Workspace licence. The full plan comparison is in our guide to Google AI Pro vs Google AI Ultra, and the choice between options inside the app is covered in our Gemini model picker guide.
··········
[[ADPLUS_300x250_1]]
GEMINI API AND GOOGLE AI STUDIO MODEL LIST WITH CODES AND STATUS.
Google AI Studio and the Gemini API expose the widest set of models. The text and multimodal models below are the ones a developer would choose for chat, coding or agents, with the status and free-tier availability shown on Google's model and pricing pages.
........
Model | Code | Status | Free tier |
Gemini 4 Carbon | gemini-4-carbon | Paid tier, launched Nov 10, 2026 | No |
Gemini 4 Argon | gemini-4-argon | General availability since Oct 13, 2026 | No |
Gemini 4 Hydrogen | gemini-4-hydrogen | Launched Nov 12, 2026 | Yes |
Gemini 3.8 Flash | gemini-3.8-flash | Stable | Yes |
Gemini 3.6 Flash | gemini-3.6-flash | Stable | Yes |
Gemini 3.5 Flash-Lite | gemini-3.5-flash-lite | Stable | Yes |
Gemini 3.1 Flash-Lite | gemini-3.1-flash-lite | Stable | Yes |
Gemini 3.1 Pro | gemini-3.1-pro-preview | Preview | No |
Gemini 3 Flash | gemini-3-flash-preview | Preview | Not stated |
Gemini 2.5 Pro | gemini-2.5-pro | Deprecated, routes to gemini-3.8-flash from Dec 1, 2026 | Yes |
Gemini 2.5 Flash | gemini-2.5-flash | Deprecated, routes to gemini-3.5-flash-lite from Dec 1, 2026 | Yes |
Gemini 2.5 Flash-Lite | gemini-2.5-flash-lite | Served until further notice, existing users only | Yes |
........
The same model list includes the specialized families. For real-time voice, Gemini 3.8 Live is the default Live API model, with a Gemini 3.8 Live Extended Thinking variant for harder reasoning; Gemini 3.1 Flash Live is marked as legacy. Text-to-speech runs on Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS, speech-to-text on Gemini 3.5 Transcribe, and real-time translation on the Gemini 3.5 Live Translate preview. Image generation and editing use Nano Banana 2.1 (gemini-nano-banana-2.1), Nano Banana 2, Nano Banana 2 Lite and Nano Banana Pro, while Gemini Omni Flash generates and edits video clips of up to 15 seconds with native audio. Google also lists Gemini Embedding 2 for multimodal embeddings, Gemini Decision 1 for the Decisions API, the Gemini Deep Research and Deep Research Max agents, the Antigravity Agent and the Gemini Robotics ER 2 preview.
Google's October 8, 2026 release notes deprecated gemini-3.7-flash, which now routes automatically to gemini-3.8-flash, and gemini-3.5-flash, which routes to gemini-3.6-flash. Gemini 2.0 Flash, Gemini 2.0 Flash-Lite, Gemini 3 Pro Preview and the 3.1 Flash-Lite preview are shut down, so code that still calls those IDs needs a new model name. Exact IDs, aliases and retirement handling are covered in our guide to Claude, Gemini and DeepSeek API model names.
··········
GEMINI MODELS ON GOOGLE CLOUD'S GEMINI ENTERPRISE AGENT PLATFORM.
Google Cloud's documentation now publishes the enterprise model catalog under the name Gemini Enterprise Agent Platform, on the pages that used to carry the Vertex AI brand. The page, last updated on November 16, 2026, lists Gemini 4 Hydrogen, Gemini 4 Carbon, Gemini 4 Argon, Gemini 3.8 Flash, Gemini 3.7 Flash, Gemini 3.6 Flash, Gemini 3.5 Flash, Gemini 3.5 Flash-Lite, Gemini 3.1 Flash-Lite, Gemini 2.5 Pro, Gemini 2.5 Flash and Gemini 2.5 Flash-Lite as generally available, together with the image models, Gemini 3.8 Live and the 3.8 TTS models. Gemini 3.1 Pro, Gemini 3 Flash, Gemini Omni 1.1 Flash and Gemini Robotics ER 2 are listed as previews.
The cloud catalog differs from the Gemini API in a few places. It still lists Gemini 3.7 Flash and 3.5 Flash as generally available, while the Gemini API has routed both to newer models since October 8. It also lists security versions that require an allowlist: Gemini 4 Carbon for cyber, Gemini 4 Argon Cyber Permissive and Gemini 3.8 Flash Cyber. Enterprise teams should check the model page for their region before migrating, because the catalog page does not show model IDs or retirement dates in its summary.
··········
GEMINI API COST EXAMPLE: THE SAME APP ON SIX MODELS.
A support assistant that handles 1,000 requests a day, with 3,000 input tokens and 800 output tokens per request, uses 90 million input tokens and 24 million output tokens over 30 days. The table applies Google's standard paid-tier prices as listed on November 17, 2026, without caching or batch discounts.
........
Model | Input cost | Output cost | Monthly total |
Gemini 4 Carbon or Argon (from Dec 1) | $360.00 | $480.00 | $840.00 |
Gemini 4 Carbon or Argon (intro, to Nov 30) | $180.00 | $240.00 | $420.00 |
Gemini 4 Hydrogen | $135.00 | $180.00 | $315.00 |
Gemini 3.8 Flash (to Dec 31) | $67.50 | $90.00 | $157.50 |
Gemini 3.8 Flash (from Jan 1, 2027) | $135.00 | $180.00 | $315.00 |
Gemini 3.5 Flash-Lite | $27.00 | $60.00 | $87.00 |
Gemini 3.1 Flash-Lite | $22.50 | $36.00 | $58.50 |
........
From January 1, 2027, Gemini 3.8 Flash and Gemini 4 Hydrogen cost exactly the same, so a team staying on 3.8 Flash only for price has until the end of December to compare the two on its own prompts. Carbon and Argon share the same standard rates, which means the choice between them comes down to the depth of reasoning a workload needs. Cached input on the Gemini 4 models is billed at 5% of the input price, with implicit caching only. All rates are in our Gemini API pricing guide.
··········
ALL GEMINI MODELS IN NOVEMBER 2026: WHAT GOOGLE CONFIRMS AND WHAT IT LEAVES OPEN.
Google's own pages confirm the structure of the lineup. Gemini 4 Carbon, Argon and Hydrogen are the current top models, with a shared 1-million-token window in the API; Gemini 3.8 Flash, 3.6 Flash, 3.5 Flash-Lite and 3.1 Flash-Lite remain the stable, cheaper options; and Gemini 2.5 Pro and 2.5 Flash stop being served as themselves on December 1, 2026, when their calls move to 3.8 Flash and 3.5 Flash-Lite. In the Gemini app, Carbon is limited to Google AI Ultra, Argon starts at Google AI Pro and Hydrogen at Google AI Plus.
Some details are still unpublished. Google does not state which model automatic selection uses for a given prompt on the free plan, the subscriptions page and the help center describe plan access differently, and the cloud catalog has not yet caught up with the Gemini API deprecations of 3.7 Flash and 3.5 Flash. Developers on Argon or Carbon face a price that doubles on December 1, and teams on Gemini 3.8 Flash face the same doubling on January 1, 2027.
··········
FOLLOW US FOR MORE.
·····
DATA STUDIOS
·····
datastudios.org




