This month has been a whirlwind for AI, with major new models arriving within days of one another. Google just released Gemini 3.8 Flash, and at first glance it has a tough job ahead of it. GPT-6 Astra is OpenAI's most capable model, while Claude Fable 5.1 is Anthropic's latest powerhouse for coding and long-running projects.
However, there's one area where Google's new Flash model has an advantage over both: video understanding.
Gemini 3.8 Flash can take video directly as an input and reason about what's happening across it. GPT-6 Astra doesn't support video input, and neither does Claude Fable 5.1. And while this might sound like a minor difference, in practice, it actually opens up an entire category of things users can ask Gemini to do that aren't native to the other two models.
Gemini can actually watch your video
Gemini 3.8 Flash accepts text, images, video, audio and PDFs as inputs, with a context window of just over 1 million tokens.
Video is particularly interesting because Google isn't simply converting a clip into a handful of screenshots and asking Gemini what's in them.
Gemini 3.8 Flash supports what Google calls "agentic video understanding." Instead of processing an entire long video in exactly the same way, the model can navigate through its timeline and decide which transcripts, frames and audio it needs to inspect to answer your question.
Google says this approach can use up to 88% fewer tokens on long-form video while delivering roughly 7% higher quality.
That means you could give Gemini a long recording and ask something surprisingly specific such as, Where did the speaker mention a particular subject? What happened immediately before someone entered the room? What was being shown on screen when a certain point was discussed?
Just from those prompts, Gemini can return answers tied to specific moments in the video rather than forcing you to scrub through it yourself.
GPT-6 Astra and Claude Fable 5.1 can't do this natively
GPT-6 Astra has a slightly larger 1.05-million-token context window and supports text and images, but OpenAI's model documentation explicitly lists both audio and video input as unsupported.
Claude Fable 5.1 has a 1-million-token context window and impressive vision capabilities, particularly for understanding diagrams, charts, tables and other visual information. But Anthropic lists its input/output capabilities as "text and images > text."
Users can extract frames, generate a transcript or use additional tools before handing that information to the model, it just isn't a native feature in the same way it is with Gemini. Gemini 3.8 Flash was designed to accept it.
Audio is part of the advantage, too
That key difference also applies to audio. Gemini 3.8 Flash accepts audio directly alongside its other input types. That means the model can reason over what it hears as part of a larger multimodal task rather than requiring you to turn everything into text first.
And video gives it both sides of the equation. Google says Gemini's video processing can work with the visual frames, audio and timestamps together. Its agentic mode can even decide whether it needs to retrieve a transcript, inspect frames or examine audio from a particular section of a video.
That's useful for everything from analyzing a lecture to finding a particular moment in a vacation video.
The takeaway
What's particularly interesting is that this isn't Google's biggest, most expensive model. Google describes Gemini 3.8 Flash as its most intelligent Flash model yet, designed to combine complex reasoning and agentic capabilities with the speed and cost efficiency the Flash family is known for.
GPT-6 Astra and Claude Fable 5.1 may be stronger choices for other jobs. Both are explicitly designed around difficult reasoning and long-running agentic work, and choosing an AI model based on a single capability would miss the bigger picture.
But video is becoming even more important of what we give AI. From class presentations and tutorials to live video rather than text, Gemini 3.8 holds the advantage over its two biggest rivals. To get started, all you have to do is upload a video or paste a link to the video in the chat and ask it to watch.
Follow Tom's Guide on Google News and add us as a preferred source to get our up-to-date news, analysis, and reviews in your feeds. Subscribe to Tom's Guide on YouTube and follow us on TikTok.
More from Tom's Guide
- AI is turning ordinary photos of kids into sexual abuse imagery — here's how to protect your family
- I asked ChatGPT, Gemini, Claude and Perplexity to rank the 10 jobs most likely to be replaced by AI — here's what they all agreed on
- Gemini just got a 'floating bubble' multitasking upgrade that's a game changer on Android phones — here's how to use it
