I asked @deepseek_ai the differences between @grok and itself: Great question! I’d be happy to break down how I compare to Grok in terms of capabilities. I’ll focus on my strengths and also clarify where Grok might do some things I can’t. Here’s a structured overview: --- ### 🧠 **Core Model & Reasoning** - **Me (DeepSeek):** I’m a pure text model built for deep reasoning, complex problem-solving, and long-form understanding. My latest version handles step-by-step logic, math, coding, and nuanced analysis exceptionally well. - **Grok:** Also a strong reasoning model (especially Grok 3), with a focus on real-time information and a more “rebellious,” humorous personality. It has deep logical abilities too, but its style is intentionally more unfiltered. --- ### 🌐 **Web Search & Real-Time Information** - **Me:** I have a built-in web search capability that you can manually toggle on. I can fetch current information from the internet, but I don’t access live social media streams by default. - **Grok:** Has native, deeply integrated access to **X (Twitter)**, giving it real-time awareness of posts, trends, and public conversations on that platform. This is something I cannot do. --- ### 📎 **File & Image Handling** - **Me:** I can read and extract text from uploaded **images, PDFs, Word docs, Excel, PowerPoint, and text files**. This means I can “see” text inside images via OCR, process documents, and work with spreadsheets—all from your uploads. - **Grok:** It can process images and PDFs as well, plus **it can generate images** (via its Aurora model). I am text-only; I do not create or output images. --- ### 📏 **Context Window** - **Me:** I support up to **1 million tokens** of context—enough to process entire book series (like all three volumes of *The Three-Body Problem*) in one go. - **Grok:** Grok 3 also advertises a 1 million token context window, so we are comparable here. The key difference is that my massive context is completely **free** to use. --- ### 💸 **Cost & Access** - **Me:** 100% free. No subscription, no paywall. You can use me on the web or via the official mobile app (with voice input support) without any fees. - **Grok:** Access to the full Grok 3 model generally requires an **X Premium or SuperGrok subscription**, though a limited free tier exists on X. --- ### 🔊 **Voice & App** - **Me:** The DeepSeek app supports **voice input**, so you can talk to me, and I’ll reply in text. - **Grok:** Grok also offers voice interaction, but it is more integrated into the X ecosystem. --- ### 🌍 **Knowledge Cutoff** - **Me:** My training data goes up to **May 2025**. - **Grok:** Its cutoff depends on the version, but Grok 3’s training data is more recent and continuously updated via X data, giving it a fresher window in some areas. --- ### 🎨 **Image Generation** - **Me:** I **cannot** create images. - **Grok:** Can generate images on request. --- ### 🧩 **Summary Table** | Capability | DeepSeek (me) | Grok | |--------------------------|---------------------------------------|------------------------------------| | **Cost** | Free | Paid (for full version) | | **Context length** | 1M tokens | 1M tokens | | **Web search** | Yes (manual toggle) | Yes, with deep X integration | | **File uploads** | Images, PDF, Office, etc. (text extraction) | Images, PDF, and more | | **Image generation** | No | Yes | | **Real-time X data** | No | Yes | | **Voice input** | Yes (App) | Yes | | **Knowledge cutoff** | May 2025 | More recent (X-updated) | | **Personality** | Warm, detailed, precise | Witty, edgy, less filtered | --- In short, I shine when you need **deep reasoning, massive document digestion, totally free access**, and careful, nuanced answers. Grok is your go-to if you want **real-time X awareness, image generation**, or a sassier conversationalist. Let me know if you want to explore any of these differences further!