Checked against the OpenRouter public models feed. Prices are copied, not estimated.

The Best Model ForWhich AI model to use, from ten labs.

The best AI model for video understanding

Ask questions about a clip you supply. This is not video generation.

Short answer

Start with Gemini 4 Neon ($3.00 in · $15.00 out per million tokens). Neon lists video input along with audio and images, in a current Gemini 4 model. Use it when the clip is evidence, not a vibe. A cheaper stop is Gemini 4 Hydrogen ($1.50 in · $7.50 out).

Quality

Gemini 4 Neon

Google (Gemini) · $3.00 in · $15.00 out

Neon lists video input along with audio and images, in a current Gemini 4 model. Use it when the clip is evidence, not a vibe.

Gemini 4 Neon is listed at $3.00 input and $15.00 output per million tokens, with a 1,048,576-token context window.

Balanced

Qwen4 Max

Alibaba (Qwen) · $1.425 in · $4.275 out

Qwen4 Max is the reviewed flagship that lists video input. Use it for a product demo or a long clip that also needs the million-token window.

Qwen4 Max is listed at $1.425 input and $4.275 output per million tokens, with a 1,000,000-token context window.

Cheapest sensible

Nova 2 Lite

Amazon (Nova) · $0.30 in · $2.50 out

Nova 2 Lite is the cheap reviewed model that lists video. Premier, despite the name, does not list video input on this feed.

Nova 2 Lite is listed at $0.30 input and $2.50 output per million tokens, with a 1,000,000-token context window.

Also on the feed

These match the factual limits of the task (modality, tools, context) and are not the reviewed picks. A row marked new has no guide note.

Video output — look not reviewed

These rows list video as an output. The Best Model For has not ranked the clips. Read the raw price fields. A $0 token price is often not the whole bill.

Questions

Can Grok 4.7 watch a video?

Not on this feed. Grok 4.7 lists text, image, and file input. Video understanding here means a row whose input modalities include video.

What if I want to generate a clip?

That is a different set of models, with output modality video. This guide does not rank how those clips look. Their prices are often not the token rates, so read the model page.