Seedance 2.5 Flash is a video generation model from ByteDance, the faster and lower-cost tier of Seedance 2.5. It supports text-to-video, image-to-video with first-frame and first-and-last-frame control, multimodal reference-to-video, video...
≈22,38 DAper second of video
+2
by ByteDance · Nov 17, 2026This model reads text and pictures and writes back text: it can take in about a thousand pages in one go and write up to 65 pages at once. It is best for sitting down with a coding job or a task where it should think first, call your own tools, and hand back a tidy form to fill in. It was built as a smaller, cheaper variant of Tencent's flagship, with about a quarter of the size, so it answers faster and costs less, and repeated use gets a discount. It cannot search the web, so jobs that need fresh pages from the internet need a model that can.
≈0,044 DAper message
↑ 14,56 DA↓ 58,45 DAper 1M tokens
+2
by Tencent · Nov 16, 2026Feed it text, pictures or video, and it writes text back, reading about 1,000 pages in one go and writing up to 390 pages in a single reply. That makes it a fit for work like reading a long report with charts and drafting answers from it. Before answering, it works through the problem step by step, can use the tools you connect, and fills forms to your exact shape. It cannot search the web on its own, so a job that needs fresh pages off the internet wants a different kind of model.
≈1,19 DAper message
↑ 474,29 DA↓ 1 422,87 DAper 1M tokens
+2
by Qwen · Nov 16, 2026It takes in text and still pictures and gives back decisions, which are answers to questions the buyer sets, each in one of three kinds the model knows. It reads about 260 pages in a single pass and writes up to 230, so a long agreement or a pile of forms can be handled in one run. It takes no sound or video and writes no finished prose, so work that needs a full report, a clip or a spoken reply is outside what it does.
≈0,030 DAper case
↑ 29,96 DAper 1M tokens
by HeyGen · Nov 13, 2026HeyGen Decision takes in written text and still pictures and gives back answers only, each to a question the buyer sets, typed and paired with a probability. It reads about 130 pages in one go and writes up to 110, so a long contract or a stack of forms can be sorted in a single request. It stops where the work turns into open writing, since it returns answers rather than prose, so a full report or a free-flowing chat needs a different kind of model.
≈0,047 DAper case
↑ 46,60 DAper 1M tokens
by HeyGen · Nov 13, 2026Recraft V4.1 Flash V2 turns a written description into pictures. It reads up to 65 pages of text at once and can return several images from one request, including tall 9:16 pictures made for phone screens. It is the fast option in the V4.1 line, so it suits quick drafts, moodboards and large batches where quick turnaround matters more than final polish. It only draws fresh pictures: it cannot edit your photos, and it does not produce see-through backgrounds or drawings that scale endlessly, so finished artwork jobs need a different kind of model.
by Recraft · Nov 13, 2026