DoubaoCatalog
Available through the NextModel gateway via Doubao.
Best forimage understanding, multimodal chat
RoutingConfigured
StreamingVision
Platform curatedNextModel gateway catalog (Go origin)
這份候選名單適合什麼用途？: Cheap LLM API
The lowest sticker price is a weak starting point. Cheap models work well for classification, short summaries, routing, and bulk drafts. They break down when the task needs careful reasoning or long agent runs. Match the model to the job first, then look at input and output rates. On this page you get price, context, capability, and provider source in one table. Run the numbers on /tools/llm-cost-calculator and /pricing before you put anything in production.
來源依據: NextModel curated catalog, provider public pricing, and OpenRouter metadata when available. · 更新日期 2026-08-05
如何使用这份名单
如何使用这份名单 (Cheap LLM API)
- 对照任务选模型. 先看「Cheap LLM API」短名单是否覆盖你的真实任务，不要只比标价。
- 用同一批提示词试跑. 挑 2–3 个候选，用业务提示词对比质量与输出长度。
- 估算月费. 用价格页或成本计算器，按预计 token 量估算月度花费。
- 定兜底与预算. 定主模型、兜底模型，设项目预算后再接生产流量。
综合价格
推薦候選 cheap llm api
先從候選名單開始，再以真實提示詞測試，並在接入正式環境路由前比較月度成本。
DoubaoCatalog
Available through the NextModel gateway via Doubao.
Best forimage understanding, multimodal chat
RoutingConfigured
StreamingVision
Platform curatedNextModel gateway catalog (Go origin)
DoubaoCatalog
Available through the NextModel gateway via Doubao.
Best forimage understanding, multimodal chat
RoutingConfigured
StreamingVision
Platform curatedNextModel gateway catalog (Go origin)
DoubaoCatalog
Available through the NextModel gateway via Doubao.
Best forimage understanding, multimodal chat
RoutingConfigured
StreamingVision
Platform curatedNextModel gateway catalog (Go origin)
比較表
按價格、提供者、上下文、能力與來源比較這份候選名單。
當你在縮小正式環境候選名單、建立備援策略或比較模型經濟性時，可使用此視圖。
| Model | Provider | Input | Output | Context | Capabilities | Best for | Latency | Status | Source |
|---|---|---|---|---|---|---|---|---|---|
| Doubao Seedance 2 0 260128doubao/doubao-seedance-2-0-260128 | Doubao | $0 / 1M tokens | $0 / 1M tokens | — | StreamingVision | image understanding, multimodal chat | 1000-3000ms | Catalog | Platform curated |
| Doubao Seedance 2 0 Fast 260128doubao/doubao-seedance-2-0-fast-260128 | Doubao | $0 / 1M tokens | $0 / 1M tokens | — | StreamingVision | image understanding, multimodal chat | 1000-3000ms | Catalog | Platform curated |
| Doubao Seedance 2 0 Mini 260615doubao/doubao-seedance-2-0-mini-260615 | Doubao | $0 / 1M tokens | $0 / 1M tokens | — | StreamingVision | image understanding, multimodal chat | 1000-3000ms | Catalog | Platform curated |
| Doubao Seedream 4 0 250828doubao/doubao-seedream-4-0-250828 | Doubao | $0 / 1M tokens | $0 / 1M tokens | — | StreamingVision | image understanding, multimodal chat | 1000-3000ms | Catalog | Platform curated |
| Doubao Seedream 4 5 251128doubao/doubao-seedream-4-5-251128 | Doubao | $0 / 1M tokens | $0 / 1M tokens | — | StreamingVision | image understanding, multimodal chat | 1000-3000ms | Catalog | Platform curated |
| Doubao Seedream 5 0 260128doubao/doubao-seedream-5-0-260128 | Doubao | $0 / 1M tokens | $0 / 1M tokens | — | StreamingVision | image understanding, multimodal chat | 1000-3000ms | Catalog | Platform curated |
| Dreamina Seedance 2 0 260128dreamina/dreamina-seedance-2-0-260128 | Dreamina | $0 / 1M tokens | $0 / 1M tokens | — | StreamingVision | image understanding, multimodal chat | 1000-3000ms | Catalog | Platform curated |
| Dreamina Seedance 2 0 Fast 260128dreamina/dreamina-seedance-2-0-fast-260128 | Dreamina | $0 / 1M tokens | $0 / 1M tokens | — | StreamingVision | image understanding, multimodal chat | 1000-3000ms | Catalog | Platform curated |
常見問題
Cheap LLM API 常見問題
What is the cheapest model in this catalog?
It depends on FX and how long answers are. In CNY production entries, Doubao Seed 2.0 Mini is usually the lowest here.
Should teams always pick the cheapest LLM API?
No. Use cheap models for repeatable, low-risk work. Keep a stronger model for hard answers, coding agents, and anything you would be embarrassed to ship wrong.
How do caching and routing lower the effective cost of a cheap LLM API?
Route easy requests to the cheap model, cache repeated prefixes when it is safe, and keep one OpenAI-compatible base URL. The calculator at /tools/llm-cost-calculator is a quick way to sanity-check the savings.