01
Start with the faster model
Use DeepSeek V4 Flash for quick questions, summaries, everyday coding, and extracted document text.
Independent multi-model AI workspace
Grok 4.5 + DeepSeek V4One workspace. Up to 1M context.
Work across code, research, images, and documents with five AI models and up to 1M tokens of context. Switch models inside one conversation and see the credit rate before you send.
Free accessUp to 1M contextText, Markdown, CSV, JSON, PDF, DOCX, JPEG, PNG, and WebP support
Try your first prompt
Your prompt moves with you to the chat workspace. Sign in is required to generate an answer.
1MMaximum context
5 modelsOne conversation
9 typesSupported files
2.5–10Credits per token
Long-context workspace
Keep the whole task in view
Up to 1M tokens of context gives complex work more room for source material, instructions, and conversation history—without splitting every task into disconnected chats.
- Document-heavy analysisCompare details across long reports and extracted documents.
- Larger coding tasksKeep more specifications, code, and decisions available together.
- Long-running researchContinue follow-ups with more of the original evidence in view.
The usable context for a request also depends on model output, supported file extraction, upload limits, and provider constraints.
What is Grok Online?
One focused workspace for DeepSeek and Grok models
Grok Online is an independent, browser-based AI workspace that brings DeepSeek V4 and selected Grok models into one conversation interface. It is built for people who want clear model choice, file-aware chat, saved history, and visible usage rates.
02
Escalate when depth matters
Move to DeepSeek V4 Pro or a Grok model for longer reasoning, complex code, and higher-output tasks.
03
Switch models inside one conversation
Choose a different model for each message while the conversation stays saved; each request includes the most recent exchanges and their supported attachments.
Model comparison
Choose the right AI model for each task
Start with the lowest-cost model that fits the job, then move up only when you need longer reasoning, code depth, or image understanding.
Free
Low latency
DeepSeek V4 Flash
Everyday Q&A, summaries, quick coding, and extracted document text
Plus / Max
Advanced tier
DeepSeek V4 Pro
Long-form reasoning, document-heavy analysis, and deeper text work
Plus / Max
Text + images
Grok 4.3 / 4.20 / 4.5
Frontier coding, difficult reasoning, and image analysis
Context is model-specific: up to 1M for DeepSeek V4, Grok 4.3, and Grok 4.20; Grok 4.5 supports 500K. Output limits also vary, and every listed model supports image analysis.
How it works
From a question to a reusable answer in three steps
The workflow stays deliberately simple: bring the task, choose the right level of reasoning, and keep the useful result in your history.
Start free chat- 01
Bring a prompt or file
Type a question or attach a supported document or image. The message and supported attachments stay saved with the conversation.
- 02
Match the model to the task
Pick Flash for speed, Pro for deeper work, or Grok for premium reasoning. Every available model accepts image input.
- 03
Continue from a saved conversation
Return to saved conversations, switch models for a follow-up, and reuse the answer in your own workflow.
Know what each answer costs
See the model's credit rate, prevent balance overdrafts, and manage paid plans through Stripe.
Independent service; not affiliated with xAI. Grok and xAI are trademarks of their respective owners.
- Five models: DeepSeek V4 Flash/Pro plus Grok 4.3, 4.20, and 4.5
- Text, Markdown, CSV, JSON, PDF, DOCX, JPEG, PNG, and WebP support
- Published per-token credit rates and conversation history
Model FAQ
DeepSeek V4 and Grok — Frequently Asked Questions
Straight answers about model access, free usage, files, subscriptions, and the relationship between Grok Online and xAI.
Which models can I use?
Grok Online now offers DeepSeek V4 Flash, DeepSeek V4 Pro, Grok 4.3, Grok 4.20, and Grok 4.5. Legacy Grok 4 and Grok 4.1 chats automatically fall back to DeepSeek V4 Flash.
Is Grok Online the official Grok website?
Grok Online is an independent platform providing access to DeepSeek and Grok models. We are not affiliated with xAI, but we provide transparent token-based billing and conversation history.
How can I use the chat for free?
Sign up for a free account to receive daily login credits. Free users can use DeepSeek V4 Flash for chat, image analysis, and extracted text from supported documents.
Which model should I choose for coding?
Use DeepSeek V4 Flash for quick edits, DeepSeek V4 Pro for deeper reasoning, and Grok 4.3, Grok 4.20, or Grok 4.5 for premium coding and synthesis. Grok uses native vision; DeepSeek images are analyzed through a vision bridge.
How are credits charged?
DeepSeek V4 Flash is 2.5 credits/token. DeepSeek V4 Pro, Grok 4.3, Grok 4.20, and Grok 4.5 are 10 credits/token. The output allowance adapts to your balance, unused pre-charge is refunded, and failed requests are fully refunded.
Can I upload files?
Yes. Text, Markdown, CSV, JSON, PDF, DOCX, JPEG, PNG, and WebP are supported. Grok analyzes images natively, DeepSeek uses a vision bridge, WebP is converted to PNG, and documents are extracted into text.
Can I switch models in one chat?
Absolutely. Toggle between DeepSeek V4 Flash, DeepSeek V4 Pro, Grok 4.3, Grok 4.20, and Grok 4.5 per message. Token usage adjusts automatically based on the model.
How do credits protect my balance?
Every request checks your balance before calling the provider and lowers the response limit when needed. It is blocked only when the balance cannot cover the input plus a minimal response, with no overdraft.
Ready when you are
Start with the fast model. Switch when the task gets harder.
Open a free account, use daily login credits with DeepSeek V4 Flash, and upgrade only when you need advanced models.
Start free chat