Gemini 3.5 Flash Lite
Sehr schnell, ideal für kurze Antworten
Best for: short questions, summaries and quick everyday answers · fast
- No login required
- Understands images
- Web search
approx. 429 tokens per response
Google · 3 models
Gemini
The Gemini models from Google are versatile and understand images too. Ask your question right here, without a Google account and without a subscription.
Genuine Google models through the official API, no account with the provider. As a guest you get 15,000 tokens with Gemini 3.5 Flash Lite. A free account adds 15,000 tokens a day, all models, uploads and saved chats.
What is Gemini?
Google Gemini is the family of AI models from Google DeepMind. It launched in December 2023, and since February 2024 it has also been the name of Google's chatbot, which was previously called Bard. The models are multimodal from the ground up, so they understand text and images together, work with a very large context, and in the Flash line are built for speed. On ChatX you use Gemini 3.5 Flash Lite, Gemini 4 Hydrogen and Gemini 4 Carbon through Google's official API, with no Google account and no subscription.
The Gemini models on ChatX
Choose the model that fits your task. Fast models are enough for everyday questions, while the larger models pay off for long texts, analyses, and code.
Gemini 4 Hydrogen
Ausgewogen mit großem Kontext
Best for: long texts, code, context-heavy analysis, images, thinking mode · balanced
- With account
- Understands images
- Web search
- Thinking mode
approx. 1,367 tokens per response
Gemini 4 Carbon
Googles Spitzenmodell für schwere Aufgaben
Best for: hard problems, code and long multi-step analysis · thorough
- With account
- Understands images
- Web search
- Thinking mode
approx. 1,823 tokens per response
New at Gemini
Blog →
Gemini 4 Carbon Is Now Available on ChatX
Google’s most capable model, Gemini 4 Carbon, replaces Gemini 4 Argon on ChatX. We show what Carbon does better, where it leads in the benc…
Nano Banana 2.1 Is Better and Costs Less
Google’s image model Nano Banana 2.1 has replaced its predecessor on ChatX and costs 8,000 tokens per image instead of 14,000. We share our…
Gemini Omni 1.2 Flash Creates Videos up to 15 Seconds Long
Google’s video model Gemini Omni 1.2 Flash has arrived in the ChatX video generator. It creates clips of 3 to 15 seconds with sound. We exp…
Use Gemini without a Google account
At Google itself, you need a Google account to use Gemini, and the Gemini app links your questions to that account. On ChatX, Gemini runs through Google's paid API: you type your question, ChatX sends it to Google and shows you the answer. You don't need a Google account for this, and as a guest you don't even need a ChatX account.
You get the current Gemini models with their strengths: very fast answers, a large context for long documents and, with an account, image understanding and web search. What you don't get are Google's own connections such as Gmail, Calendar or Drive. In return, Gemini sits right next to ChatGPT, Claude, Grok and DeepSeek here, and you can switch in the middle of a conversation.
You can also create images on ChatX with Google models, just not in the chat itself: in the AI image generator you'll find Google's Nano Banana models alongside GPT Image from OpenAI. Videos with sound are made in the AI video generator with Gemini Omni 1.2 Flash, Google's video model.
What you can do with Gemini on ChatX
Write texts
Emails, posts and job applications in your tone, written especially fast with Gemini.
Summarize
Paste long texts and get them down to a few sentences; with an account you can also upload PDF and Word files.
Research current topics
For Gemini models with the globe icon, you turn on web search and get current sources with the answer.
Understand images
With an account you upload photos, screenshots or charts and have Gemini 4 Hydrogen explain what they show.
Code
Have code explained, bugs found and small functions written; according to Google, Gemini 4 Hydrogen is built for fast, efficient everyday coding.
Learn and understand
Get difficult topics explained step by step with examples, and with thinking mode even multi-step math problems.
Gemini compared with the other models on ChatX
Pick Gemini when you need speed, are working with a lot of text or want an image explained with your account; Gemini 3.5 Flash Lite is among the low-cost models on ChatX, Gemini 4 Hydrogen is the all-rounder built for speed, and for hard problems there is Gemini 4 Carbon, Google's top model. ChatGPT and Claude often write more polished language in long, carefully crafted texts, and Claude is especially strong with code and structured documents, Grok shines on current topics, and DeepSeek is very cheap for logic and math problems but runs on servers in China. If an answer doesn't fit, just switch to another model in the same chat.
| Model | Strength | No sign-up | Features |
|---|---|---|---|
| ChatGPT | All-rounder for writing, analysis and code | GPT-6 Luna | Web search · Understands images · Thinking mode |
| Claude | Long texts, careful language, code | Claude Haiku 5.5 | Web search · Understands images · Thinking mode |
| Gemini | Fast, large context, understands images | Gemini 3.5 Flash Lite | Web search · Understands images · Thinking mode |
| Grok | Direct tone, quick answers | Grok 4.3 | Understands images · Thinking mode |
| DeepSeek | Math, logic and code at a low price | DeepSeek V4.1 Flash | Web search · Understands images · Thinking mode |
All models run in the same chat. Switch any time without losing your history.
Google Gemini, explained simply
Gemini is Google's family of AI models. The Flash models are built for short response times, handle images as well as text and search the web on request. On ChatX you use the Gemini models through the official API without a Google account, with an English interface and transparent billing per reply. Gemini 3.5 Flash Lite runs without signing up and is suited to short questions and summaries; with a free ChatX account you also get Gemini 4 Hydrogen with thinking mode for longer texts, code and analyses and Gemini 4 Carbon, Google's top model for hard problems, and you can upload images and documents and save your chats.
It's this easy
-
1
Enter your question
Type your question or task into the input field. You don't need to log in for this.
-
2
Choose a model
Start with the default model or switch to a more powerful one anytime.
-
3
Keep going
With a free account, you save your chats, upload files, and get free tokens every day.
Costs and free allowance
- Without signing up, you get 15,000 tokens to try it out. That's enough for around 34 answers with Gemini 3.5 Flash Lite.
- With a free account, you get 15,000 tokens as a gift every day.
- If you need more, you can buy tokens with no subscription or choose a monthly plan. Billing is based on the model's actual usage.
Privacy: what happens to my input?
Your request goes from ChatX to Google's official Gemini API, and the answer comes back the same way. You don't need a Google account for this, and you never sign in to Google.
ChatX stores your chats only in your own account, and you can delete them at any time. ChatX uses the paid Gemini API. According to Google's terms for it (as of November 2026), Google does not use these inputs and answers to improve its products and logs them only briefly to prevent abuse; when web search is turned on, Google keeps the requests for 30 days.
Don't enter sensitive information such as passwords, health data or other people's personal data. You'll find the details in the privacy policy.
Frequently asked questions
What happened to Bard?
Bard was Google's chatbot, which got the first Gemini models at the end of 2023. On February 8, 2024, Google renamed Bard to Gemini, and since then the chatbot and the model family share the same name. On ChatX you use the current Gemini models from this family.
Do I need a Google account?
No. On ChatX you use Gemini without a Google account and start without signing up. For Gemini 4 Hydrogen, Gemini 4 Carbon and uploads, a free ChatX account is all you need.
Which Gemini model should I choose?
For short questions, Gemini 3.5 Flash Lite, which the page starts with, is enough. For longer texts, context-heavy analysis or thinking mode, switch to Gemini 4 Hydrogen with a free account. For the toughest tasks, such as code or multi-step analysis, you will also find Gemini 4 Carbon there, Google's top model.
Can I create images with Gemini?
Not with the chat model. You create images on ChatX with the AI image generator, even without signing up, including with Google's Nano Banana models. For existing photos there are separate tools such as the AI photo editor or the Background remover. Videos with sound are made in the AI video generator with an account, including with Gemini Omni 1.2 Flash.
Can Gemini search the internet?
For the Gemini models with web search (globe icon in the input field), you turn on search in the chat. The answer then includes current sources.
Does Gemini answer in English?
Yes. Gemini understands and writes English and many other languages, and it answers in the language you ask in. If a reply does come back in another language, just ask for English in the chat.
Can I upload images and documents?
Yes, with a free ChatX account: you upload photos, screenshots, PDF and Word files and ask Gemini 4 Hydrogen or Gemini 4 Carbon about them. Without signing up, uploads are not enabled.
What does Gemini cost on ChatX?
Gemini 3.5 Flash Lite is among the low-cost models on ChatX, Gemini 4 Hydrogen uses somewhat more tokens and the top model Gemini 4 Carbon more still; you can see the approximate usage per answer in the model overview. Without signing up you try it with a starting allowance, and after that you only pay for actual usage per answer, with no mandatory subscription.
The history of Google Gemini
From the merger of Google Brain and DeepMind in spring 2023, through the renaming of the Bard chatbot, to Gemini 4 Argon, Carbon and Hydrogen in fall 2026: in just over three years, Gemini became Google's central AI brand for models, app and search.
Newest events first
-
Gemini 4 · 2026
-
Modell
Gemini 4 Hydrogen: cutting-edge intelligence for everyone
Hydrogen has been generally available in the Gemini API since November 12, 2026, including the free tier. Google wants it to bring “cutting-edge intelligence to everyone” and built it for speed and efficiency in everyday coding and agentic workflows; it joins Carbon, the strongest thinking model, and Argon in the Gemini 4 lineup. Hydrogen handles up to 1 million tokens of context, understands text, images, audio, video and PDF, and offers four thinking levels in the API, from low to xhigh; it costs $1.50 per million input tokens and $7.50 per million output tokens. On ChatX, Gemini 4 Hydrogen has replaced Gemini 3.8 Flash as the all-rounder since November 13, 2026, with a free account, thinking mode, web search and image upload.
1M token contextWidely available since Nov 12On ChatX
-
Modell
Gemini 4 Carbon: Google's strongest model for everyone
Carbon has been generally available in the Gemini API since November 10, 2026. Google calls it its strongest model for sustained coding work, multimodal knowledge work, scientific reasoning and cyber defense; Argon stays on as the faster, leaner sibling. Carbon thinks longer per answer but costs the same per token. On ChatX, Gemini 4 Carbon has replaced Argon as Google's top model since November 11, 2026, with a free account, thinking levels, web search and image upload.
Strongest Google modelWidely available since Nov 10On ChatX
-
Modell
Gemini Omni 1.2 Flash: video from a conversation
Google releases Omni 1.2 Flash in the Gemini API and in Gemini for Business. The model creates clips up to 15 seconds long, can be refined step by step in conversation and accepts reference images. All output carries the SynthID watermark. On ChatX you use Omni 1.2 Flash in the AI video generator: clips of 3 to 15 seconds with sound, from text, a first and last frame or reference images, in up to 4K.
15 s video with soundReference imagesOn ChatX
-
Modell
Gemini 4 Carbon announced
Google announces Carbon as its most powerful model yet for long-running engineering and security tasks. At first, only US government agencies and selected cyber defenders test it; the broad launch follows on November 10, 2026.
Most powerful Google modelEarly testing onlyBroad launch on Nov 10
-
Produkt
Veo 3.1 leaves the Gemini API
Google shuts down the Veo 3.1 video models in the Gemini API and points to Gemini Omni as the replacement, so video generation in the API now runs on the Omni line. Veo has not been selectable on ChatX since then; the AI video generator offers Gemini Omni 1.2 Flash instead.
Shut down on Oct 22Replaced by Gemini OmniNo longer on ChatX
-
Modell
Nano Banana 2.1: better images at half the price
Google releases Nano Banana 2.1, the successor to Nano Banana 2, built on Gemini 3.6 Flash. The image model improves on image quality, prompt adherence, text in images and infographics, keeps characters consistent across multiple edits, accepts up to 14 reference images and delivers panoramic formats up to 8:1 and resolutions up to 4K. In the API, an image costs half as much, and Nano Banana 2 has been marked as deprecated since then. In the Gemini app, Google calls it its best model for image generation and editing and offers it only with a Google AI subscription. On ChatX, Nano Banana 2.1 replaces its predecessor in the AI image generator and works with a free account.
Half the API priceUp to 4K and 8:1On ChatX
-
Modell
Gemini 4 Argon: long tasks in one go
Argon kicks off the fourth generation. The model is built for long, multi-step work, for example in code and in financial and legal questions, and can output up to one million tokens in one go. After a staged rollout with cyber defenders and a voluntary review by the US government, Argon has been widely available in the Gemini API, app, Search and Antigravity since October 13, 2026; in the app, Pro and Ultra subscribers get it. On ChatX, Gemini 4 Carbon replaced Argon on November 11, 2026.
1M token outputWidely available since Oct 13Replaced on ChatX
-
Gemini 3 · 2025/26
-
Modell
Gemini 3.8 Flash
After 3.6 and 3.7 Flash in the summer, the Flash line gets another step up in September with stronger autonomous coding. Google highlights its resistance to prompt injection and safeguards against cyber and CBRN misuse; the 3.8 Flash Cyber variant goes only to selected defenders through the Fairwind program. On ChatX, Gemini 4 Hydrogen replaced Gemini 3.8 Flash on November 13, 2026.
Agentic codingPrompt injection protectionReplaced on ChatX
-
Regulierung
EU code for transparent AI content
Google signs the EU AI Act code of practice on labeling AI-generated content and relies on SynthID to do so. To build shared watermarking standards, Google works with Apple, ElevenLabs, NVIDIA and OpenAI, among others.
EU AI ActSynthIDIndustry standard
-
Modell
Gemini 3.6 Flash operates the computer, 3.5 Flash-Lite arrives
Gemini 3.6 Flash brings computer use from the specialist models into the general-purpose model: it can click, scroll and fill in interfaces. At the same time, 3.5 Flash-Lite arrives as a fast budget option and 3.5 Flash Cyber for selected defenders, while 3.5 Pro stays in partner testing. On ChatX, Gemini 3.5 Flash Lite is the guest model you use without signing up.
Computer useFlash-LiteGuest model on ChatX
-
Produkt
I/O 2026: Gemini 3.5 Flash, Omni and Agent Spark
At I/O, Google shows Gemini 3.5 Flash as a faster model that acts more on its own, for example in coding and tool use; 3.5 Pro is announced for later. Gemini Omni combines multimodal understanding with image and video generation, and the Gemini app gets Spark, a personal agent built on 3.5 and Antigravity. At this point, Google cites more than 900 million monthly users of the app.
3.5 FlashGemini Omni900M users
-
Modell
Gemini 3.1: Pro, Nano Banana 2 and Flash-Lite
In February, Gemini 3.1 Pro raises core intelligence and ships right away in the app, the API, NotebookLM and Antigravity. Shortly after, Nano Banana 2 follows with better image quality, cleaner text rendering and image search to back up facts. In March come 3.1 Flash-Lite, the fastest and cheapest 3.x variant with selectable thinking levels, and 3.1 Flash Live for real-time voice with SynthID in the audio.
3.1 ProNano Banana 2Flash-Lite with thinking levels
-
Modell
Gemini 3 launches
The third generation focuses on stronger reasoning, multimodality and autonomous coding and ships immediately in the Gemini app, AI Studio, Vertex AI, Search and the new development environment Antigravity. Two days later, Nano Banana Pro arrives for images with better text rendering. In December, Deep Think for Ultra subscribers and Gemini 3 Flash for fast, low-cost tasks follow.
Agentic codingAntigravityNano Banana Pro
-
Gemini 2.5 · 2025
-
Forschung
Gemini 2.5 Computer Use
A preview model built on 2.5 Pro operates graphical interfaces in the browser: clicking, scrolling, typing. For risky steps, it asks a human to confirm. With this, Google moves from simple tool calls to an agent that operates websites by itself.
Browser agentPreviewConfirmation for risky steps
-
Regulierung
Google signs the EU code for AI models
Google says it will sign the voluntary code of practice for general-purpose AI models under the EU AI Act, but raises concerns about copyright and trade secrets. The European Commission then lists Google as a signatory. For users in Europe, this is the most important regulatory milestone around Gemini.
EU AI ActVoluntary codeWith reservations
-
Modell
Gemini 2.5 Pro and Flash become standard
From June, both core models of the 2.5 generation are considered stable and run in production; 2.5 Flash-Lite starts as a preview and in July becomes the family's cheapest tier, with up to one million tokens of context and optional thinking. From August, Ultra subscribers get Deep Think, which pursues several solution paths in parallel; according to Google, a special version reached gold medal level at the International Mathematical Olympiad.
2.5 stableFlash-LiteDeep Think
-
Produkt
Google AI Pro and Google AI Ultra
Gemini Advanced becomes Google AI Pro, and above it Google introduces the much pricier Ultra plan with the highest limits and early access to new features. At I/O, Google also shows Deep Think mode, native audio output, summaries of the thinking steps, MCP support and the browser agent Project Mariner.
Free / Pro / UltraDeep Think shownProject Mariner
-
Modell
Gemini 2.5: thinking moves to the core
With 2.5 Pro Experimental, Google introduces a generation that reasons before answering instead of offering thinking only as an add-on; multimodality and a one-million-token context remain. In April, 2.5 Flash follows as the family's first hybrid model: developers turn thinking on or off and set a computing budget. The thinking mode you use with Gemini 4 Hydrogen and Gemini 4 Carbon on ChatX comes from this line.
Thinking model1M token contextThinking budget
-
Gemini 2.0 · 2024/25
-
Firma
New AI principles, expanded safety framework
Google revises its AI Principles: weapons and certain surveillance applications are no longer listed as blanket exclusions, which draws criticism from employees and observers. At the same time, DeepMind expands its Frontier Safety Framework with security protocols and the risk of deceptive models. A day later, 2.0 Flash becomes generally available, joined by 2.0 Pro, Flash-Lite and Flash Thinking with an explicit thinking mode.
AI Principles changedFrontier Safety v22.0 family complete
-
Modell
Gemini 2.0 Flash and Deep Research
With 2.0, Google ushers in the “agentic era”: 2.0 Flash understands text, images and audio, outputs text, images and speech itself, calls tools like Search directly and is faster than 1.5 Pro. The app launches Deep Research, which plans multi-step web research, goes through sources and writes a report. The new Live API enables real-time voice and video conversations that you can interrupt.
Native toolsDeep ResearchLive API
-
Gemini 1.5 · 2024
-
Produkt
Gemini Live and Gems
Gemini Live brings natural voice conversations you can interrupt, with ten voices, first for Advanced subscribers on Android and in English; since October 2024, Live has also spoken German. With Gems you build your own specialized assistants, and Imagen 3 improves image generation in the app.
Voice conversationsGerman since Oct 2024Gems
-
Produkt
Gemini app in Germany, two million tokens for developers
The Gemini app officially launches in Germany. For developers, 1.5 Pro and 1.5 Flash become stable and available as paid models at the end of May; at the end of June, Google opens the two-million-token context window broadly and adds code execution and context caching. This makes Gemini especially suited to very long documents, codebases and videos.
App in Germany2M tokensContext caching
-
Modell
Gemini 1.5 Flash and Project Astra
At I/O, Google presents 1.5 Flash, a smaller model built for speed and low cost; it becomes the Flash line you use on ChatX today. Gemini Advanced switches to 1.5 Pro with a one-million-token context and document upload. As a look ahead, Project Astra shows an assistant that sees and talks in real time.
Flash line is born1M tokens in the appProject Astra
-
Regulierung
Election questions restricted worldwide
As a precaution against misinformation, Gemini only answers questions about elections in a limited way worldwide. In the same month, DeepMind publishes its first systematic tests of dangerous capabilities such as deception and cyberattacks; in May, the Frontier Safety Framework follows, with fixed thresholds that trigger additional safeguards.
ElectionsDangerous capabilities testedFrontier Safety v1
-
Firma
Images of people paused
After viral examples of historically implausible images of people, Google temporarily stops Gemini from generating people. Google admits that overcorrecting for diversity and overly cautious filters led to wrong and sometimes offensive results. The case remains the app's best-known early bias controversy.
Bias controversyImage feature pausedFeb 22/23, 2024
-
Modell
Gemini 1.5 Pro: one million tokens
Gemini 1.5 Pro is the first Gemini to use a mixture-of-experts architecture and, according to Google, reaches the level of 1.0 Ultra in many areas. The standard context is 128,000 tokens, and selected testers get one million: enough for entire codebases, long audio recordings or videos.
Mixture of experts1M tokensLong context
-
Produkt
Bard becomes Gemini
Google renames the Bard chatbot to Gemini, launches the Gemini mobile app and introduces Gemini Advanced, a subscription that is the first to unlock Ultra 1.0. A week earlier, Bard with Gemini Pro had expanded to more than 40 languages and more than 230 countries, including Germany, Austria and Switzerland. From now on, Gemini is Google's central consumer AI brand.
Bard renamedGemini AdvancedUltra 1.0
-
Beginnings · 2023
-
Modell
Gemini 1.0: Ultra, Pro and Nano
Google releases Gemini 1.0 as a natively multimodal model family in three sizes: Ultra for the hardest tasks, Pro for broad use and Nano for smartphones. Bard gets Gemini Pro right away, the Pixel 8 Pro gets Nano, and developers get Pro on December 13 through AI Studio and Vertex AI. The demo video draws criticism: it looks like a smooth live conversation but was stitched together from still frames and text prompts and sped up.
Three sizesBard with Gemini ProDemo controversy
-
Firma
Gemini is announced
A few weeks after Google Brain and DeepMind merged into Google DeepMind on April 20, 2023, Sundar Pichai mentions the name Gemini for the first time at I/O: a model still in training, multimodal from the ground up, in several sizes and designed with tool use, planning and memory in mind. It is the starting point of the family that later replaces the Bard chatbot.
Google DeepMind foundedI/O 2023Multimodal from the ground up
Compiled from public announcements by Google and press reports; updated when new models arrive.
More AI models on ChatX
Last reviewed: 13. November 2026 · Source: Google