← Brands

StreamLake

StreamLake

Multi-model API provider · Usage-based billing

StreamLake is Kuaishou's AI cloud platform. Its Vanchin model service provides an OpenAI-compatible API for proprietary KAT-Coder models and third-party DeepSeek, Qwen, GLM, Kimi, MiniMax and MiMo models, billed per million tokens in USD with published cache pricing. Image and video generation models are priced per image or per second; availability and prices follow the official Vanchin documentation.

·

Usage-based API; see model prices.

Listed models · 31

Distinct model versions are listed once; aliases are not additional models. Prices below are for API usage.

Qwen3.5-397B-A17B

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

Qwen3.5-397B-A17B

Input / output: text, image, video → text

Model developer: Alibaba Qwen · Service provider: StreamLake · Market: com

Model family: Qwen · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 0.6
OutputUSD 3.6
Cached readUSD 0.12

StreamLake (Vanchin) provides the hosted model API; model developer: Alibaba Qwen. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 262,144 tokens

Maximum output: 65,536 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: Alibaba Qwen. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Qwen3.6-35B-A3B

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

Qwen3.6-35B-A3B

Input / output: text, image, video → text

Model developer: Alibaba Qwen · Service provider: StreamLake · Market: com

Model family: Qwen · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 0.375
OutputUSD 2.25
Cached readUnverified

StreamLake (Vanchin) provides the hosted model API; model developer: Alibaba Qwen. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 262,144 tokens

Maximum output: 65,536 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: Alibaba Qwen. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Qwen3.6-27B

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

Qwen3.6-27B

Input / output: text, image, video → text

Model developer: Alibaba Qwen · Service provider: StreamLake · Market: com

Model family: Qwen · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 0.6
OutputUSD 3.6
Cached readUnverified

StreamLake (Vanchin) provides the hosted model API; model developer: Alibaba Qwen. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 262,144 tokens

Maximum output: 65,536 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: Alibaba Qwen. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Qwen3.8-2.4T-A95B

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

Qwen3.8-2.4T-A95B

Input / output: text, image, video → text

Model developer: Alibaba Qwen · Service provider: StreamLake · Market: com

Model family: Qwen · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 2
OutputUSD 6
Cached readUSD 0.17

StreamLake (Vanchin) provides the hosted model API; model developer: Alibaba Qwen. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 1,048,576 tokens

Maximum output: 131,072 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: Alibaba Qwen. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Qwen3.8-27B

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

Qwen3.8-27B

Input / output: text, image, video → text

Model developer: Alibaba Qwen · Service provider: StreamLake · Market: com

Model family: Qwen · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 0.5
OutputUSD 3
Cached readUSD 0.05

StreamLake (Vanchin) provides the hosted model API; model developer: Alibaba Qwen. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 1,048,576 tokens

Maximum output: 131,072 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: Alibaba Qwen. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

DeepSeek-R1-250528

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

DeepSeek-R1-250528

Input / output: text → text

Model developer: DeepSeek · Service provider: StreamLake · Market: com

Model family: DeepSeek · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 0.571
OutputUSD 2.286
Cached readUSD 0.229

StreamLake (Vanchin) provides the hosted model API; model developer: DeepSeek. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 131,072 tokens

Maximum output: 16,384 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: DeepSeek. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

DeepSeek-V3

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

DeepSeek-V3

Input / output: text → text

Model developer: DeepSeek · Service provider: StreamLake · Market: com

Model family: DeepSeek · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 0.286
OutputUSD 1.143
Cached readUSD 0.114

StreamLake (Vanchin) provides the hosted model API; model developer: DeepSeek. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 65,536 tokens

Maximum output: 8,192 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: DeepSeek. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

DeepSeek-V4-Flash

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

DeepSeek-V4-Flash

Input / output: text → text

Model developer: DeepSeek · Service provider: StreamLake · Market: com

Model family: DeepSeek · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 0.14
OutputUSD 0.28
Cached readUSD 0.028

StreamLake (Vanchin) provides the hosted model API; model developer: DeepSeek. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 1,048,576 tokens

Maximum output: 393,216 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: DeepSeek. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

DeepSeek-V4-Flash-0731

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

DeepSeek-V4-Flash-0731

Input / output: text → text

Model developer: DeepSeek · Service provider: StreamLake · Market: com

Model family: DeepSeek · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 0.44
OutputUSD 1.32
Cached readUSD 0.014

StreamLake (Vanchin) provides the hosted model API; model developer: DeepSeek. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 1,048,576 tokens

Maximum output: 393,216 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: DeepSeek. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

DeepSeek-V4-Pro

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

DeepSeek-V4-Pro

Input / output: text → text

Model developer: DeepSeek · Service provider: StreamLake · Market: com

Model family: DeepSeek · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 1.74
OutputUSD 3.48
Cached readUSD 0.145

StreamLake (Vanchin) provides the hosted model API; model developer: DeepSeek. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 1,048,576 tokens

Maximum output: 393,216 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: DeepSeek. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

DeepSeek-V4-Pro-0813

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

DeepSeek-V4-Pro-0813

Input / output: text → text

Model developer: DeepSeek · Service provider: StreamLake · Market: com

Model family: DeepSeek · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 1.32
OutputUSD 3.96
Cached readUSD 0.044

StreamLake (Vanchin) provides the hosted model API; model developer: DeepSeek. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 1,048,576 tokens

Maximum output: 393,216 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: DeepSeek. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

DeepSeek-V4.1-Flash

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

DeepSeek-V4.1-Flash

Input / output: text, image → text

Model developer: DeepSeek · Service provider: StreamLake · Market: com

Model family: DeepSeek · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 0.3
OutputUSD 1.2
Cached readUSD 0.006

StreamLake (Vanchin) provides the hosted model API; model developer: DeepSeek. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 1,048,576 tokens

Maximum output: 393,216 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: DeepSeek. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

DeepSeek-V4.1-Flash-1017

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

DeepSeek-V4.1-Flash-1017

Input / output: text, image → text

Model developer: DeepSeek · Service provider: StreamLake · Market: com

Model family: DeepSeek · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 0.3
OutputUSD 1.2
Cached readUSD 0.006

StreamLake (Vanchin) provides the hosted model API; model developer: DeepSeek. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 1,048,576 tokens

Maximum output: 393,216 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: DeepSeek. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

GLM-5

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

GLM-5

Input / output: text → text

Model developer: Z.ai · Service provider: StreamLake · Market: com

Model family: GLM · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 1
OutputUSD 3.2
Cached readUSD 0.2

StreamLake (Vanchin) provides the hosted model API; model developer: Z.ai. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 204,800 tokens

Maximum output: 131,072 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: Z.ai. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

GLM-5.1

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

GLM-5.1

Input / output: text → text

Model developer: Z.ai · Service provider: StreamLake · Market: com

Model family: GLM · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 1.4
OutputUSD 4.4
Cached readUSD 0.26

StreamLake (Vanchin) provides the hosted model API; model developer: Z.ai. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 204,800 tokens

Maximum output: 131,072 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: Z.ai. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

GLM-5.2

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

GLM-5.2

Input / output: text → text

Model developer: Z.ai · Service provider: StreamLake · Market: com

Model family: GLM · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 1.4
OutputUSD 4.4
Cached readUSD 0.26

StreamLake (Vanchin) provides the hosted model API; model developer: Z.ai. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 1,048,576 tokens

Maximum output: 131,072 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: Z.ai. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

GLM-5.3

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

GLM-5.3

Input / output: text → text

Model developer: Z.ai · Service provider: StreamLake · Market: com

Model family: GLM · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 1.4
OutputUSD 4.4
Cached readUSD 0.26

StreamLake (Vanchin) provides the hosted model API; model developer: Z.ai. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 1,048,576 tokens

Maximum output: 131,072 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: Z.ai. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

GLM-5.3-Flash

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

GLM-5.3-Flash

Input / output: text, image, video → text

Model developer: Z.ai · Service provider: StreamLake · Market: com

Model family: GLM · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 0.15
OutputUSD 0.5
Cached readUSD 0.03

StreamLake (Vanchin) provides the hosted model API; model developer: Z.ai. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 1,048,576 tokens

Maximum output: 131,072 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: Z.ai. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Kimi-K2-Instruct-0905

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

Kimi-K2-Instruct-0905

Input / output: text → text

Model developer: Moonshot AI · Service provider: StreamLake · Market: com

Model family: Kimi · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 0.571
OutputUSD 2.286
Cached readUSD 0.143

StreamLake (Vanchin) provides the hosted model API; model developer: Moonshot AI. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 262,144 tokens

Maximum output: 32,768 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: Moonshot AI. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Kimi K2.6

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

Kimi K2.6

Input / output: text, image, video → text

Model developer: Moonshot AI · Service provider: StreamLake · Market: com

Model family: Kimi · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 0.95
OutputUSD 4
Cached readUSD 0.16

StreamLake (Vanchin) provides the hosted model API; model developer: Moonshot AI. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 262,144 tokens

Maximum output: 262,144 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: Moonshot AI. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Kimi-K2.7-Code

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

Kimi-K2.7-Code

Input / output: text, image, video → text

Model developer: Moonshot AI · Service provider: StreamLake · Market: com

Model family: Kimi · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 0.95
OutputUSD 4
Cached readUSD 0.19

StreamLake (Vanchin) provides the hosted model API; model developer: Moonshot AI. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 262,144 tokens

Maximum output: 32,768 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: Moonshot AI. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

MiniMax-M2.5

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

MiniMax-M2.5

Input / output: text → text

Model developer: MiniMax · Service provider: StreamLake · Market: com

Model family: MiniMax · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 0.3
OutputUSD 1.2
Cached readUSD 0.03

StreamLake (Vanchin) provides the hosted model API; model developer: MiniMax. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 204,800 tokens

Maximum output: 131,072 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: MiniMax. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

MiniMax-M3

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

MiniMax-M3

Input / output: text, image, video → text

Model developer: MiniMax · Service provider: StreamLake · Market: com

Model family: MiniMax · Currency: USD

  • Input: USD 0.6 / 1,000,000 tokens · Input length ≤ 512K tokens
  • Output: USD 2.4 / 1,000,000 tokens · Input length ≤ 512K tokens
  • Cache hit: USD 0.12 / 1,000,000 tokens · Input length ≤ 512K tokens
  • Input: USD 1.2 / 1,000,000 tokens · Input length > 512K tokens
  • Output: USD 4.8 / 1,000,000 tokens · Input length > 512K tokens
  • Cache hit: USD 0.24 / 1,000,000 tokens · Input length > 512K tokens

Complete billing tiers (context or time conditions: [{"id":"streamlake-minimax-m3:tier:1","condition":"Input length ≤ 512K tokens","min_tokens":0,"max_tokens":512000,"output_min_tokens":null,"output_max_tokens":null,"lowerInclusive":true,"upperExclusive":true,"currency":"USD","unit":"tokens","denominator":1000000,"input_price":0.6,"output_price":2.4,"cache_read_price":0.12,"cache_write_price":null,"cache_storage_price":null,"components":[{"dimension":"input","amount":"0.6","currency":"USD","unit":"tokens","denominator":1000000,"condition":"Input length ≤ 512K tokens","published":true,"free":false,"tierId":"streamlake-minimax-m3:tier:1"},{"dimension":"output","amount":"2.4","currency":"USD","unit":"tokens","denominator":1000000,"condition":"Input length ≤ 512K tokens","published":true,"free":false,"tierId":"streamlake-minimax-m3:tier:1"},{"dimension":"cache_hit","amount":"0.12","currency":"USD","unit":"tokens","denominator":1000000,"condition":"Input length ≤ 512K tokens","published":true,"free":false,"tierId":"streamlake-minimax-m3:tier:1"}]},{"id":"streamlake-minimax-m3:tier:2","condition":"Input length > 512K tokens","min_tokens":512000,"max_tokens":null,"output_min_tokens":null,"output_max_tokens":null,"lowerInclusive":true,"upperExclusive":true,"currency":"USD","unit":"tokens","denominator":1000000,"input_price":1.2,"output_price":4.8,"cache_read_price":0.24,"cache_write_price":null,"cache_storage_price":null,"components":[{"dimension":"input","amount":"1.2","currency":"USD","unit":"tokens","denominator":1000000,"condition":"Input length > 512K tokens","published":true,"free":false,"tierId":"streamlake-minimax-m3:tier:2"},{"dimension":"output","amount":"4.8","currency":"USD","unit":"tokens","denominator":1000000,"condition":"Input length > 512K tokens","published":true,"free":false,"tierId":"streamlake-minimax-m3:tier:2"},{"dimension":"cache_hit","amount":"0.24","currency":"USD","unit":"tokens","denominator":1000000,"condition":"Input length > 512K tokens","published":true,"free":false,"tierId":"streamlake-minimax-m3:tier:2"}]}]

StreamLake (Vanchin) provides the hosted model API; model developer: MiniMax. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 1,048,576 tokens

Maximum output: 524,288 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: MiniMax. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

MiMo-V2.6-Pro

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

MiMo-V2.6-Pro

Input / output: text → text

Model developer: Xiaomi MiMo · Service provider: StreamLake · Market: com

Model family: MiMo · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 0.435
OutputUSD 0.87
Cached readUSD 0.0036

StreamLake (Vanchin) provides the hosted model API; model developer: Xiaomi MiMo. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 1,048,576 tokens

Maximum output: 131,072 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: Xiaomi MiMo. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

MiMo-V2.6-Flash

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

MiMo-V2.6-Flash

Input / output: text → text

Model developer: Xiaomi MiMo · Service provider: StreamLake · Market: com

Model family: MiMo · Currency: USD

Rates per million tokens
Price dimensionCurrent rate
InputUSD 0.14
OutputUSD 0.28
Cached readUSD 0.0028

StreamLake (Vanchin) provides the hosted model API; model developer: Xiaomi MiMo. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

Documented API context: 1,048,576 tokens

Maximum output: 131,072 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

StreamLake (Vanchin) provides the hosted model API; model developer: Xiaomi MiMo. Billed per million tokens in USD with published cache pricing. The official pricing document does not publish a raw API model ID, so use the model identifier shown in the console.

KAT-Coder-Pro V2.5

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

KAT-Coder-Pro V2.5

Input / output: text → text

Model developer: Kuaishou · Service provider: StreamLake · Market: com

Model family: KAT-Coder · Model catalog only; see service-channel quotes

KAT-Coder is StreamLake's proprietary coding model family. The official Vanchin model catalog lists this version, but the official pay-as-you-go pricing document does not publish a token price. The historical Coding Plan page returns HTTP 404 and no current purchase entry was verified, so the model is kept as catalog-only and is not presented as a purchasable API offer.

Documented API context: 262,144 tokens

Maximum output: 81,920 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

KAT-Coder is StreamLake's proprietary coding model family. The official Vanchin model catalog lists this version, but the official pay-as-you-go pricing document does not publish a token price. The historical Coding Plan page returns HTTP 404 and no current purchase entry was verified, so the model is kept as catalog-only and is not presented as a purchasable API offer.

KAT-Coder-Air V2.5

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

KAT-Coder-Air V2.5

Input / output: text → text

Model developer: Kuaishou · Service provider: StreamLake · Market: com

Model family: KAT-Coder · Model catalog only; see service-channel quotes

KAT-Coder is StreamLake's proprietary coding model family. The official Vanchin model catalog lists this version, but the official pay-as-you-go pricing document does not publish a token price. The historical Coding Plan page returns HTTP 404 and no current purchase entry was verified, so the model is kept as catalog-only and is not presented as a purchasable API offer.

Documented API context: 262,144 tokens

Maximum output: 81,920 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

KAT-Coder is StreamLake's proprietary coding model family. The official Vanchin model catalog lists this version, but the official pay-as-you-go pricing document does not publish a token price. The historical Coding Plan page returns HTTP 404 and no current purchase entry was verified, so the model is kept as catalog-only and is not presented as a purchasable API offer.

KAT-Coder-Pro V1

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

KAT-Coder-Pro V1

Input / output: text → text

Model developer: Kuaishou · Service provider: StreamLake · Market: com

Model family: KAT-Coder · Model catalog only; see service-channel quotes

KAT-Coder is StreamLake's proprietary coding model family. The official Vanchin model catalog lists this version, but the official pay-as-you-go pricing document does not publish a token price. The historical Coding Plan page returns HTTP 404 and no current purchase entry was verified, so the model is kept as catalog-only and is not presented as a purchasable API offer.

Documented API context: 262,144 tokens

Maximum output: 32,768 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

KAT-Coder is StreamLake's proprietary coding model family. The official Vanchin model catalog lists this version, but the official pay-as-you-go pricing document does not publish a token price. The historical Coding Plan page returns HTTP 404 and no current purchase entry was verified, so the model is kept as catalog-only and is not presented as a purchasable API offer.

KAT-Coder-Pro V2

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

KAT-Coder-Pro V2

Input / output: text → text

Model developer: Kuaishou · Service provider: StreamLake · Market: com

Model family: KAT-Coder · Model catalog only; see service-channel quotes

KAT-Coder is StreamLake's proprietary coding model family. The official Vanchin model catalog lists this version, but the official pay-as-you-go pricing document does not publish a token price. The historical Coding Plan page returns HTTP 404 and no current purchase entry was verified, so the model is kept as catalog-only and is not presented as a purchasable API offer.

Documented API context: 204,800 tokens

Maximum output: 81,920 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

KAT-Coder is StreamLake's proprietary coding model family. The official Vanchin model catalog lists this version, but the official pay-as-you-go pricing document does not publish a token price. The historical Coding Plan page returns HTTP 404 and no current purchase entry was verified, so the model is kept as catalog-only and is not presented as a purchasable API offer.

KAT-Coder-Exp-72B-1010

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

KAT-Coder-Exp-72B-1010 · legacy · Historical record

KAT-Coder-Exp-72B-1010

Input / output: text → text

Model developer: Kuaishou · Service provider: StreamLake · Market: com

Model family: KAT-Coder · Model catalog only; see service-channel quotes

KAT-Coder is StreamLake's proprietary coding model family. The official Vanchin model catalog lists this version, but the official pay-as-you-go pricing document does not publish a token price. The historical Coding Plan page returns HTTP 404 and no current purchase entry was verified, so the model is kept as catalog-only and is not presented as a purchasable API offer. The official model catalog marks this experimental version as retired.

Documented API context: 131,072 tokens

Maximum output: 32,768 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

KAT-Coder is StreamLake's proprietary coding model family. The official Vanchin model catalog lists this version, but the official pay-as-you-go pricing document does not publish a token price. The historical Coding Plan page returns HTTP 404 and no current purchase entry was verified, so the model is kept as catalog-only and is not presented as a purchasable API offer. The official model catalog marks this experimental version as retired.

KAT-Coder-Air V1

API models and prices · 1

Usage-based billing: text rates are usually per million tokens; other units and tiers are stated per entry. Compare only the same currency, unit and billing conditions. Unverified cache pricing does not mean free.

KAT-Coder-Air V1 · legacy · Historical record

KAT-Coder-Air V1

Input / output: text → text

Model developer: Kuaishou · Service provider: StreamLake · Market: com

Model family: KAT-Coder · Model catalog only; see service-channel quotes

KAT-Coder is StreamLake's proprietary coding model family. The official Vanchin model catalog lists this version, but the official pay-as-you-go pricing document does not publish a token price. The historical Coding Plan page returns HTTP 404 and no current purchase entry was verified, so the model is kept as catalog-only and is not presented as a purchasable API offer. The official model catalog marks this lightweight version as retired.

Documented API context: 131,072 tokens

Maximum output: 32,768 tokens

Requires a StreamLake account, Vanchin model service activation and an API key; use the model identifier shown in the console.

Official source · Last verified: 2026-11-08T00:00:00.000Z

Capabilities and usage limits

KAT-Coder is StreamLake's proprietary coding model family. The official Vanchin model catalog lists this version, but the official pay-as-you-go pricing document does not publish a token price. The historical Coding Plan page returns HTTP 404 and no current purchase entry was verified, so the model is kept as catalog-only and is not presented as a purchasable API offer. The official model catalog marks this lightweight version as retired.