Model Catalog and Pricing
LaoZhang API currently offers 198 models. Find a model ID, price, token group, and endpoint by use case or vendor. Data updated 2026-10-18, 09:14 (UTC+8); call logs show the final charge.
Choose a model by use case
Narrow the list by what you need to build, then follow the guide for the request format. Counts are models online now.
Recently added
Models that first appeared in the LaoZhang API public pricing in the last 60 days, newest first. Launch notes are in Announcements.Prices by vendor
Each vendor has one table per use case, with newer versions first. Wide tables scroll sideways on phones.
OpenAI 89
Text and chat 65
Token group for these models: default.
Image generation and editing 15
Embeddings 3
Token group for these models: default.
Speech and audio 6
Token group for these models: default.
Anthropic 11
Text and chat
Token group for these models: claude_code, default.
Google 21
Text and chat 14
Token group for these models: default.
Image generation and editing 7
Token group for these models: GPTImage2 Sora2 Enterprise (×1.2), default.
xAI 17
Text and chat 14
Token group for these models: default.
Image generation and editing 3
Token group for these models: default.
DeepSeek 4
Text and chat
Token group for these models: default.
Alibaba 32
Text and chat 25
Token group for these models: default.
Video generation 6
Token group for these models: Wan (×0.1714).
Embeddings 1
Token group for these models: default.
ByteDance 13
Text and chat 4
Token group for these models: default.
Image generation and editing 5
Token group for these models: default.
Video generation 4
Token group for these models: SeeDance2 (×0.1714).
Zhipu AI 3
Text and chat
Moonshot 1
Text and chat
Token group for these models: default.
MiniMax 1
Text and chat
Token group for these models: default.
Black Forest Labs 5
Image generation and editing
Token group for these models: default.
StepFun 1
Text and chat
Token group for these models: default.
Tiered pricing
Models marked † are priced by the token count of each request, and the whole request is billed at one tier. Catalog tables show the first tier; all tiers are below. Call logs show the tier and charge actually applied.
How pricing works
- All prices are in US dollars. Usage-based input, output, and cache-read prices are per 1M tokens; per-call models are priced per call.
- Tables show list prices. Account contracts or dedicated routes can be priced differently; the console shows what applies to you.
- The cache-read price applies only when a request hits the model’s prompt cache. “—” means no cache-read price is configured; cache misses are billed at the normal input price.
- A “×” value after a token group is the multiplier applied to the listed price when you call through that group.
- Video models are billed per task or by video length; see the Wan and Seedance 2 guides.
gpt-image-2has both per-call and usage-based routes; the GPT Image 2 group guide explains which billing mode your token needs.- GPT Image official-forward routes are billed by tokens: the Sora2Official group matches OpenAI’s official price, and the GPTImage2 Sora2 Enterprise group is the official price plus 20%.
- The Endpoints column shows the protocol entry points in the pricing configuration. “OpenAI-compatible” means the matching OpenAI-format endpoint (text, images, audio, or moderation depending on the model); it does not mean every client parameter is supported.
Enterprise purchasing and discounts
For enterprise purchasing, volume usage, contracts, invoices, discounts, or special payment terms, contact the site owner or support. The documentation does not publish fixed discount thresholds or percentages; final terms follow the confirmed arrangement, the console, and call logs.
Frequently asked questions
Are these prices live?
The page is generated from the system pricing configuration and shows when it was last updated. Price changes, account groups, or dedicated contracts can make the console price differ; the console and call logs are final.
How do I see which models my token can call?
Send GET https://api.laozhang.ai/v1/models with that token. The response lists the models the token can call right now, and the list depends on the token group.
What if a model is not in this catalog?
The catalog lists online models in the public pricing configuration only. A missing model is either not offered publicly or has been retired; launches and retirements are posted in Announcements. For dedicated enterprise routes, contact support.
Which protocols do Claude models support?
Claude models are listed again from October 18, 2026, and accept both the Anthropic Messages and the OpenAI-compatible format. Output is capped at 32,768 tokens per request, and PDF document input and Anthropic server tools such as web_search are not available. See the Claude protocol page for usage and the full list of limits.
How does tiered pricing work?
The token count of each request selects a tier, and the whole request is billed at that tier’s input, output, and cache-read prices. See the tiered pricing table above.
I found the model ID. How do I call it?
Check the endpoint and token group in its table, then create a token in that group in the console. Text models use /v1/chat/completions for the OpenAI-compatible route; image, video, audio, and other models follow the guides listed under “Choose a model by use case”.
How do I arrange enterprise purchasing, volume usage, or discounts?
Contact the site owner or support with your account, models, expected usage, and contract and invoice needs. The documentation does not promise a fixed discount threshold or percentage.
Pricing snapshot generated: 2026-10-18T01:14:54.035Z