ByteDance Seed
Four 2026 priorities are world models (internal goal of a release by year-end benchmarked against Genie 3), keeping Seedance video near the top, a stronger coding foundation, and commercializing Doubao (200M+ daily users reported) for office work.
- Founded
- 2023
- Headquarters
- Beijing, China
- Now
- Entered the world-model race late; insiders say it trailed state of the art by about 10% in early 2026 (KrASIA citing 36Kr). Seedance 2.0 reportedly drives a large share of AI revenue.
Key people
- Wu Yonghui (head of Seed)
- Zhou Chang (world models)
- Fan Haoqi (ex-Meta FAIR, new world-model group)
Latest launches
- SeedRealtime 5 August 2026
Native audio-visual full-duplex LLM that fuses audio, video and text in one stream and can speak up unprompted; rolled out at scale at launch. - Seedance 2.5 ByteDance · 31 July 2026
Seedance 2.5 makes 30-second single-pass clips with synced audio, up to 50 multimodal references, region-level edits and 4K, previewed 2026-06-23. - Seed 2.1 (Pro, Turbo) 23 June 2026
Agent-first successor to Seed 2.0 in Pro and Turbo sizes, claiming the highest GDPVal score and best MobileWorld GUI result. - Seed 2.0 Lite (0428) 28 April 2026
Late-April Seed2.0 Lite update became the first omni-modal understanding model in the Seed family, adding native audio input alongside video, image and text. - Seedance 2.0 studio cease-and-desists ByteDance · 16 February 2026
Disney, Netflix, Paramount, Warner Bros. and Sony sent cease-and-desist letters over Seedance 2.0; the MPA called it 'systemic infringement'. - Seed 2.0 (Pro, Lite, Mini, Code) 14 February 2026
Flagship Doubao generation in Pro, Lite, Mini and Code models; Pro claims IMO, CMO and ICPC gold results and token prices about ten times lower. - Seedance 2.0 ByteDance · 12 February 2026
Seedance 2.0 jointly generates audio and video from text, image, audio and video references (9+3+3 per prompt), and Hollywood objected to it. - Stable-DiffCoder 15 January 2026
8B open diffusion code LLM that beats comparable autoregressive and diffusion code models, using block-diffusion continued pretraining on the Seed-Coder pipeline.
Launches by year
22 launches and papers since 15 May 2024, oldest first within each year.
- 2026 8
- Stable-DiffCoder, Seedance 2.0, Seed 2.0 (Pro, Lite, Mini, Code), Seedance 2.0 studio cease-and-desists, Seed 2.0 Lite (0428), Seed 2.1 (Pro, Turbo), Seedance 2.5, SeedRealtime
- 2025 13
- UI-TARS, Doubao-1.5-pro, OmniHuman-1, DAPO, Seed1.5-Thinking, Seedream 3.0, Seed1.5-VL, Seed-Coder-8B, Seedance 1.0, Seed Diffusion Preview, Seed-OSS-36B, Seedream 4.0, Seedance 1.5 Pro
- 2024 1
- Doubao large model family
Sources
- kr-asia.com/bytedance-sets-four-ai-priorities-for-2026
- seed.bytedance.com/en/blog/seedream-4-0-officially-released-beyond-drawing-into-imaginatio
- arxiv.org/abs/2509.20427
- technode.com/2026/02/14/bytedance-releases-doubao-seed-2-0-positions-pro-model-against-gpt
- en.wikipedia.org/wiki/Seedance_2.0
- kr-asia.com/bytedance-raises-volcano-engines-maas-revenue-target-on-seedance-2-0-growth
- eu.36kr.com/en/p/3736438995907079
This profile was checked and corrected against its sources on 6 October 2026. How we check