Vendor · Language models · Image models · Video models
ByteDance has 12 models in the directory: 8 language models, 2 image models and 2 video models.
In the directory: 8 language models, 2 image models and 2 video models. The vendor publishes open weights for none of them.
Every model there ever was. Models with a profile lead to every benchmark, price and source.
| Seed 2.1 Turbo | Jun 23, 2026 | API only | n/a | $0.50 / $2.50 | n/a | 256K | 17 |
| Seed 2.0 Code | Feb 14, 2026 | API only | n/a | $0.50 / $3 | n/a | 256K | 16 |
| Seed 2.0 Lite | Feb 14, 2026 | Deprecated | n/a | $0.25 / $2 | n/a | 256K | 14 |
| Seed 2.0 Mini | Feb 14, 2026 | Deprecated | n/a | $0.10 / $0.40 | n/a | 256K | 13 |
| Seed 2.0 Pro | Feb 14, 2026 | API only | n/a | $0.50 / $3 | n/a | 256K | 15 |
| Seed 1.8 | Dec 18, 2025 | Deprecated | n/a | $0.25 / $2 | n/a | 256K | 17 |
| Seed 1.6 | Jun 15, 2025 | Deprecated | n/a | $0.25 / $2 | n/a | 256K | 13 |
| Seed 1.6 Flash | Jun 15, 2025 | Deprecated | n/a | $0.075 / $0.30 | n/a | 256K | 9 |
| Seedream 5.0 Pro | July 2026 | Active | $0.09 per image | 33 |
| Seedream 4.0 | September 2025 | Active | n/a | 24 |
| Dreamina Seedance 2.5 720p | Jul 31, 2026 | Active | n/a | 12 |
| Dreamina Seedance 2.0 720p | March 2026 | Active | $9.07 per minute | 30 |
Video model · Jul 31, 2026
New in Seedance 2.5 are 30-second audio-video clips, multiple extension rounds, and smoother transitions between scenes. A single run can use up to 30 images, 10 video clips, and 10 audio clips as references.
Image model · July 2026
Seedream 5.0 Pro is documented as its own BytePlus ModelArk endpoint with 1K, 1.5K, and 2K tiers. Its Pro ID runs separately from Seedream 4.0, and a higher version number does not automatically make prompt, reference, or resolution options compatible.
Language model · Jun 23, 2026
Seed 2.1 Turbo puts ByteDance’s text, image, and video inputs into an accelerated API serving form with 256K tokens of context. Turbo therefore names the serving form of the Seed 2.1 line, not a released local checkpoint.
Video model · March 2026
What is distinctive about Seedance 2.0 is its unified audio-video architecture for text, images, audio, and video. References can guide performance, lighting, shadows, and camera movement; the 720p profile generates up to 15 seconds.
Language model · Feb 14, 2026
Seed 2.0 Code is ByteDance’s specialized product role for programming work. The API model is listed with 256K tokens of context, multimodal understanding, and tool use in TRAE and through the Volcano interface. Its parameter count and weights are not public, so the name does not identify a local checkpoint.
Language model · Feb 14, 2026 · Deprecated
Seed 2.0 Lite is ByteDance’s cost-oriented service tier for frequent multimodal calls. The proprietary API model handles text, images, and video with 256K tokens of context. Because ByteDance publishes neither a parameter count nor weights, Lite describes a serving tier rather than a local model size.
Language model · Feb 14, 2026 · Deprecated
Seed 2.0 Mini is the smallest listed service tier in the Seed 2.0 family. ByteDance publishes no parameter count. Text, image, and video inputs use a 256K-token API context; no local weight package is provided.
Language model · Feb 14, 2026
Seed 2.0 Pro is ByteDance’s large proprietary tier for visual puzzles, documents, video, and other unstructured information. It handles text, images, and video with 256K tokens of context. The parameter count is unpublished, with access provided through ByteDance and BytePlus APIs.
Language model · Dec 18, 2025 · Deprecated
Seed 1.8 combines three reasoning modes with image understanding and a 256K-token context. ByteDance presents the proprietary service as a generalized agent model for search, coding, and GUI tasks. ByteDance publishes no parameter count, so the accessible model form is defined by its product role and tool surface.
Image model · September 2025
Seedream 4.0 combines text-to-image generation and editing in one model. ByteDance documents text, image, and combined inputs, single- and multi-image editing, up to a dozen references, and 4K output. Identity, structure, and style can therefore change together in one scene.
Language model · Jun 15, 2025 · Deprecated
Seed 1.6 is ByteDance’s multimodal deep-thinking model with 230B total parameters, 23B active parameters, and 256K tokens of context. Its technical description highlights adaptive thinking time, AdaCoT, and GUI-agent use. The model can therefore vary the amount of internal planning between a direct answer and a more involved task.
Language model · Jun 15, 2025 · Deprecated
Seed 1.6 Flash is ByteDance’s compact API variant of its multimodal deep-thinking model. It accepts text, images, and video, supports up to 256K tokens of context, and can answer directly with thinking disabled. The parameter count is unpublished, and no weights are provided.
| Vendor | Models | Model types | Best language model | Top 10 | Open weights | New in 2026 |
|---|---|---|---|---|---|---|
| Anthropic | 33 | 33 language models | Claude Opus 5.5 (Rank 1 to 3) | 4 | 0 | 13 |
| OpenAI | 63 | 46 language models, 4 image models, 2 video models, 9 audio models, 2 specialist models | GPT-6 Astra (Rank 4 to 8) | 2 | 4 | 27 |
| 66 | 43 language models, 4 image models, 6 video models, 11 audio models, 2 specialist models | Gemini 4 Carbon (Rank 1 to 5) | 2 | 19 | 26 | |
| Meta | 32 | 29 language models, 1 image model, 1 video model, 1 specialist model | Muse Wonder Full (Rank 6 to 9) | 2 | 26 | 7 |
| DeepSeek | 23 | 23 language models | DeepSeek V4.1 Flash 1017 (Rank 13 to 20) | 0 | 23 | 9 |
| Moonshot AI | 12 | 12 language models | Kimi K3.1 Pro (Rank 10 to 13) | 0 | 10 | 7 |
| xAI | 16 | 12 language models, 2 image models, 2 video models | Grok 4.7 (Rank 18 to 24) | 0 | 1 | 12 |
| ByteDance | 12 | 8 language models, 2 image models, 2 video models | not ranked yet | 0 | 0 | 8 |
By release date. Grey: retired models.
8 releases since January 2026, by release date.
Prices per 1M tokens for the current language models, according to ByteDance or the provider linked in each model profile.
| Current model | Input | Output | Context |
|---|---|---|---|
| Seed 1.6 FlashDeprecated | $0.075 | $0.30 | 256K |
| Seed 2.0 MiniDeprecated | $0.10 | $0.40 | 256K |
| Seed 2.0 LiteDeprecated | $0.25 | $2 | 256K |
| Seed 1.8Deprecated | $0.25 | $2 | 256K |
| Seed 1.6Deprecated | $0.25 | $2 | 256K |
| Seed 2.1 Turbo | $0.50 | $2.50 | 256K |
| Seed 2.0 Code | $0.50 | $3 | 256K |
| Seed 2.0 Pro | $0.50 | $3 | 256K |
| Current model | Price |
|---|---|
| Seedream 5.0 Pro | $0.09 per image |
| Dreamina Seedance 2.0 720p | $9.07 per minute |
The directory lists 12 ByteDance models: 8 language models, 2 image models and 2 video models. It counts every model ByteDance ever released, including retired ones. 12 are currently available, the newest being Dreamina Seedance 2.5 720p, Seedream 5.0 Pro and Seed 2.1 Turbo.
Dreamina Seedance 2.5 720p, released on Jul 31, 2026. The timeline below shows every release by date.
Through the API per 1M tokens (input / output): Seed 1.6 Flash $0.075 / $0.30. The prices of all 8 current language models are in the price table.
No. ByteDance publishes open weights for none of the 12 models; they only run through the API, its own apps or partners.
Up to 256,000 tokens (≈ 380 A4 pages), which 8 of the current models reach, including Seed 2.1 Turbo. Every model's value is in the model list.
Data as of Nov 1, 2026