In partnership with

I asked Perplexity if any of the Foundational LLM (as compared to special tools) can create Gifs of 2 to 10 frames like the one above and this is what it told me:

First I asked it:

What Foundational AI Among the Big 5 of Chatgpt, Claude, Gemini, Grok and Perplexity can generate animated gifs of say 2 to 10 frames

This is what it told me:

The "big five" chatbots—ChatGPT (OpenAI), Claude (Anthropic), Gemini (Google), Grok (xAI), and Perplexity—are all advanced large language models (LLMs) with multimodal capabilities, but as of mid-2025, none of them natively generate animated GIFs (such as 2–10 frame animations) directly from a text prompt within their standard chat interfaces. Here’s a breakdown of their current generative capabilities and limitations regarding animated GIF creation:

ChatGPT (OpenAI)

• ChatGPT (including GPT-4o and GPT-4.5) is highly capable in text, code, and image generation, with some versions supporting image creation via DALL-E integration1.

• While ChatGPT can generate static images and code for simple animations (like SVG or HTML/CSS/JS), it does not directly output animated GIFs from text prompts in its chat interface.

• Users can, however, ask ChatGPT to generate a sequence of images or code to create frame-by-frame animations, which can then be assembled into a GIF using external tools1.

Claude (Anthropic)

• Claude 3.7 is a powerful LLM with advanced reasoning, code generation, and vision analysis2.

• It can help generate code or scripts for creating animations and can describe how to use external tools or libraries to assemble GIFs, but does not natively generate animated GIFs from prompts2.

Gemini (Google)

• Gemini 2.5 is Google’s most advanced AI, excelling in reasoning, coding, and multimodal tasks3.

• While Gemini can generate and analyze images, and may assist in creating animation code or workflows, it does not directly output animated GIFs from a prompt in its chat interface3.

Grok (xAI)

• Grok 3 (and the upcoming Grok 3.5) features image generation and analysis, real-time web search, and advanced reasoning4.

• It can generate images and possibly help with code for animations, but there is no indication that Grok can natively create animated GIFs from a text prompt within its chat interface4.

Perplexity

• Perplexity AI is a conversational search engine powered by LLMs (GPT-4, Claude-2) and excels at providing information, summarizing documents, and generating text-based content5.

• It does not natively generate images or animations, though it can guide users to tools or provide code to create GIFs externally5.

Summary Table

Conclusion

While the "big five" AI chatbots are highly capable across text, code, and image generation, none currently offer direct, native generation of animated GIFs (2–10 frames) from text prompts in their chat interfaces. Users typically need to use specialized generative AI tools (such as AnimateDiff, ToonCrafter, or GifGPT) or follow code-based workflows (which these chatbots can help generate) to create animated GIFs.

Add to follow-up

Check sources

1. https://www.inc.com/ben-sherry/an-entrepreneurs-guide-to-every-chatgpt-ai-model/91177567

2. https://pwskills.com/blog/what-is-claude-ai-model/

3. https://blog.google/technology/google-deepmind/gemini-model-thinking-updates-march-2025/

4. https://builtin.com/articles/grok

5. https://www.flyrank.com/blogs/ai-insights/what-model-does-perplexity-ai-use-an-in-depth-exploration

6. https://teamai.com/blog/large-language-models-llms/understanding-different-chatgpt-models/

7. https://openai.com/index/introducing-gpt-4-5/

8. https://model-spec.openai.com

9. https://platform.openai.com/docs/models

10. https://anthemcreation.com/en/artificial-intelligence/choose-best-chatgpt-model-projects-2025-complete-guide-comparative/

OK, then what tools can do it?

Several AI tools can generate animated GIFs from as few as 2 to 10 frames, leveraging advances in generative AI, diffusion models, and animation-specific adapters.

Here’s an overview of the most relevant options available as of 2025:

AI Models and Tools for Animated GIF Generation

1. AnimateDiff (Stable Diffusion + Motion Adapter)

• AnimateDiff is a prominent open-source solution that extends Stable Diffusion with a motion adapter trained on video clips. It generates short animations (including GIFs) from text prompts or initial images.

• You can specify the number of frames (typically 2–16, with 2–10 being well-supported), but higher frame counts require more GPU memory.

• The workflow involves providing a prompt, negative prompt (optional), and frame count; the model then produces a sequence of images that can be compiled into a GIF.

• Hardware requirements are significant (ideally a modern GPU with at least 8–16GB VRAM)1.

2. ToonCrafter

• ToonCrafter is a specialized tool that can interpolate smooth animations from as few as two input frames.

• The process involves creating two key poses, stylizing them with Stable Diffusion (optionally using ControlNet for pose consistency), and then letting ToonCrafter generate the in-between frames to form a short animation or GIF.

• This approach is highly accessible for generating stable, short animations from minimal input, but also benefits from substantial VRAM for best results3.

3. GifGPT

• GifGPT uses DALL-E to generate a spritesheet (e.g., 12 frames in a 3x4 grid) based on a text prompt, then slices and assembles the frames into an animated GIF.

• It’s designed for quick, theme-based GIF generation directly from text, with the number of frames typically set by the tool (commonly 12, but potentially configurable)4.

4. Online AI GIF Generators (e.g., Picsart, Mage Space, Gifer, Simplified)

• These platforms offer user-friendly interfaces to generate GIFs from text, images, or videos, often leveraging proprietary or open-source generative models.

• Many allow you to set the number of frames or duration, and some (like Picsart) focus on text-to-GIF workflows for rapid, creative output28.

• They’re accessible to non-technical users and suitable for quick, visually appealing results, though customization and control over frame-by-frame content may be limited compared to AnimateDiff or ToonCrafter.

Comparison Table

Summary

• AnimateDiff and ToonCrafter are foundational AI models capable of generating animated GIFs from as few as 2 frames, offering high control and quality for technical users13.

• GifGPT and commercial platforms like Picsart provide quick, accessible text-to-GIF generation, suitable for most creative needs with minimal technical setup248.

• For hands-on workflows and maximum control over animation details, AnimateDiff and ToonCrafter are recommended. For rapid, creative GIFs from prompts, GifGPT and online generators are ideal.

These tools collectively represent the current state-of-the-art in foundational AI for short, animated GIF generation from minimal input.

1. https://www.youtube.com/watch?v=Qspqx6ZFs9c

2. https://www.videoproc.com/resource/ai-gif-generator.htm

3. https://www.reddit.com/r/sdforall/comments/1dfo9cf/ai_animation_from_only_2_frames/

4. https://theee.ai/tools/gifgpt-ZxX4Dtgn

5. https://www.aibase.com/tool/14003

6. https://mockey.ai/blog/ai-gif-generator/

7. https://leonardo.ai/ai-video-generator/

8. https://picsart.com/ai-gif-generator/

9. https://developer.nvidia.com/blog/new-foundational-models-and-training-capabilities-with-nvidia-tao-5-5/

10. https://www.toolify.ai/ai-news/create-stunning-animated-gifs-with-photoshop-79898

Here’s one review of the top 4 paid chatbots

Here is another take on the subject

Veo3 from Google Gemini is making fantastic movies.

when you’re ready to move on from keyboard mashing the chatbots, the next step could be n8n - an outstanding workflow automation tool that can orchestrate LLM and Agents in a step by step manner

Here's a free 8 hour course in n8n, but it takes a bit of concentration and rewatching to get it. Nate Hock (below) is good but I’m looking for alternatives.

Reply to this email (yes, you can do that) if you’re interested in a tutorial on how to use n8n to create custom document chatbots and agents of your own material similar to Google’s NotebookLM.

Let me know if you would like to get involved in any projects and what you would like to see in this news letter

There’s so much going on , i’m thinking of doing another post this week

Best regards

• Frank,

• [email protected] 

The #1 AI Newsletter for Business Leaders

Join 400,000+ executives and professionals who trust The AI Report for daily, practical AI updates.

Built for business—not engineers—this newsletter delivers expert prompts, real-world use cases, and decision-ready insights.

No hype. No jargon. Just results.

Subscribe free—trusted by leaders