Doubao Models by ByteDance
About
The Doubao large language model (LLM) family by ByteDance is a suite of cutting-edge AI foundation models designed to advance LLM technology. They focus on overcoming key challenges such as model self-learning, memory capabilities, long-text generation, and interpretability 1. The lineup includes Doubao Pro, offering a 128k context length with fine-tuning options, and Doubao Lite, which optimizes for lower token costs and improved latency 1. Doubao Character enables personalized character creation with enhanced context-awareness. These models support numerous applications across text, speech, audio, music, and video generation and are utilized by over 50 ByteDance businesses, also being available to external enterprises through Volcano Engine 5. Additionally, the Doubao team is proactive in research and publications, showcasing their drive for innovation in AI 9.
Current Variants
Use-when guidance is derived from seed capabilities, context, release, and replacement fields.
Use when the workload needs 4k context and 36B parameters.
Use when the workload needs 128k context and 128B parameters.
Use when the workload needs 32k context and 32B parameters.
Use when the workload needs 32k context and 32B parameters.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Seed-OSS 36B Instruct | Use when the workload needs 4k context and 36B parameters. | 2025-06 | 4k context36B parameters | Current |
| Doubao Pro 256K | Use when the workload needs 256k context. | 2024-05 | 256k context | Current |
| Doubao Lite 128K | Use when the workload needs 128k context. | 2024-05 | 128k context | Current |
| Doubao Pro 128K | Use when the workload needs 128k context and 128B parameters. | 2024-05 | 128k context128B parameters | Current |
| Doubao Pro 32K | Use when the workload needs 32k context and 32B parameters. | 2024-05 | 32k context32B parameters | Current |
| Doubao Lite 32K | Use when the workload needs 32k context and 32B parameters. | 2024-05 | 32k context32B parameters | Current |
| Doubao | Use when provider availability and model metadata match the workload. | 2023-08 | — | Current |
Release Timeline
3 release groupsSpecifications(7 models)
| Model | Released | Context | Parameters |
|---|---|---|---|
| Seed-OSS 36B Instruct | 2025-06 | 4k | 36B |
| Doubao Pro 256K | 2024-05 | 256k | — |
| Doubao Lite 128K | 2024-05 | 128k | — |
| Doubao Pro 128K | 2024-05 | 128k | 128B |
| Doubao Pro 32K | 2024-05 | 32k | 32B |
| Doubao Lite 32K | 2024-05 | 32k | 32B |
| Doubao | 2023-08 | — | — |
Available From(2 providers)
Pricing
| Model | Provider | Input / 1M | Output / 1M | Type |
|---|---|---|---|---|
| Doubao Lite 32K | Volcengine | $0.044 | $0.088 | Serverless |
| Doubao Lite 128K | Volcengine | $0.118 | $0.147 | Serverless |
| Doubao Pro 32K | Volcengine | $0.118 | $0.294 | Serverless |
| Doubao Pro 256K | Volcengine | $0.735 | $1.324 | Serverless |
| Doubao Pro 128K | Volcengine | $0.735 | $1.324 | Serverless |
Frequently Asked Questions
- What is Doubao used for?
- Doubao is used for math-heavy prompts and long-context generation. The family description and listed model capabilities point to those workloads as the best fit.
- How does Doubao compare to Seed?
- Doubao by ByteDance is strongest where you need math-heavy prompts, while Seed by ByteDance is the closest related family to check for vision and multimodal work. Doubao has 7 listed variants and reaches up to 256k context, while Seed reaches up to 256k context, so compare the specs and pricing tables before choosing a production model.
- Which Doubao model should I use?
- For the lowest listed input price, start with Doubao Lite 32K through Volcengine at $0.044/1M input tokens. For the most capable/latest local choice, evaluate Doubao Pro 256K with 256k context.
