Alibaba Cloud
Qwen3.6 Plus
Alibaba Tongyi Qwen
qwen3.6-plus Qwen's balanced flagship — covers general Q&A, research reading, coding, and multimodal tasks.
- Context
- 1M tokens
- Pricing Tier
- ¥2 input / ¥12 output
- Modalities
- Text / Image
← Back to Quick Start
Domestic LLM Overview · Updated 2026-04-28
Domestic LLM & API Channel Guide
A quick reference for key parameters, pricing, and access channels for mainstream Chinese LLMs — helping Aether users choose and configure models efficiently.
Domestic LLM Overview
Key parameters and official links for commonly used Chinese models; actual pricing and availability per vendor console.
DeepSeek
DeepSeek V4 Pro
DeepSeek
deepseek-v4-pro DeepSeek's top-tier model, currently at 2.5× discount; V4 Flash offers a cost-effective alternative.
- Context
- 1M tokens
- Pricing Tier
- Pro ¥12 input / ¥24 output · Flash ¥1 input / ¥2 output
- Modalities
- Text
Moonshot AI
Kimi K2.6
Moonshot AI Kimi
kimi-k2.6 A next-gen model for long-horizon coding, Agent workflows, and multimodal understanding — strong Chinese performance.
- Context
- 256K tokens
- Pricing Tier
- ¥6.5 input / ¥27 output
- Modalities
- Text / Image
Zhipu
GLM-5.1
Zhipu Z.AI
glm-5.1 Optimized for long-horizon Coding Agent and complex engineering tasks — stable structured output and tool calling.
- Context
- 200K tokens
- Pricing Tier
- ¥10 input / ¥50 output · Coding Plan
- Modalities
- Text
Xiaomi
Xiaomi MiMo-V2.5
Xiaomi MiMo
mimo-v2.5 / mimo-v2.5-pro Pro flagship for complex software engineering and long-horizon Agent tasks; standard version supports native multimodal — image, video, and audio understanding.
- Context
- 1M tokens
- Pricing Tier
- Pro ¥4 input / ¥16 output · Standard ¥1 input / ¥4 output
- Modalities
- Text / Image
MiniMax
MiniMax-M2.7
MiniMax
MiniMax-M2.7 Flagship text model with self-iteration capability; highspeed version delivers significantly faster inference at double the price.
- Context
- 1M tokens
- Pricing Tier
- ¥2.1 input / ¥8.4 output
- Modalities
- Text
Doubao
Doubao seed-2.0-pro
ByteDance Doubao
doubao-seed-2-0-pro ByteDance's flagship Agent model for complex reasoning and long-chain tasks; 1.5-pro is the cost-effective variant.
- Context
- 256K tokens
- Pricing Tier
- seed-2.0 ¥3.2 input / ¥16 output · 1.5 ¥0.8 input / ¥2 output
- Modalities
- Text / Image
Baidu
ERNIE 5.0
Baidu Wenxin
ernie-5.0 Baidu's latest native all-modal LLM — deep thinking and visual understanding capabilities fully upgraded.
- Context
- 128K tokens
- Pricing Tier
- ¥10 input / ¥40 output
- Modalities
- Text / Image
Purchase & Access Channels
Official channels are run directly by model vendors; platform channels aggregate multiple models via cloud or inference providers; subscription channels offer coding-focused bundles.
Official Channels
0Subscription Channels
0Platform Channels
0Config Reference
When filling in Aether, mainly confirm Provider type, Base URL, and Model field.
| Channel | Provider Type | Base URL | Model Example |
|---|---|---|---|
| Xiaomi MiMo | OpenAI / Anthropic Compatible | https://api.xiaomimimo.com/v1https://api.xiaomimimo.com/anthropic | mimo-v2.5mimo-v2.5-pro |
| Alibaba Cloud Bailian | OpenAI / Anthropic Compatible | https://dashscope.aliyuncs.com/compatible-mode/v1https://dashscope.aliyuncs.com/apps/anthropic | qwen3.6-plus |
| DeepSeek | OpenAI / Anthropic Compatible | https://api.deepseek.comhttps://api.deepseek.com/anthropic | deepseek-v4-prodeepseek-v4-flash |
| Kimi | Moonshot AI / OpenAI / Anthropic Compatible | https://api.moonshot.cn/v1https://api.moonshot.cn/anthropic | kimi-k2.6 |
| Zhipu | Zhipu AI / OpenAI / Anthropic Compatible | https://open.bigmodel.cn/api/paas/v4/https://open.bigmodel.cn/api/anthropic | glm-5.1 |
| MiniMax | OpenAI / Anthropic Compatible | https://api.minimaxi.com/v1https://api.minimaxi.com/anthropic | MiniMax-M2.7MiniMax-M2.7-highspeed |
| Baidu Qianfan (ERNIE) | OpenAI Compatible | https://qianfan.baidubce.com/v2 | ernie-5.0 |
| Platform Channel Config Examples | |||
| Volcengine Ark | OpenAI Compatible | https://ark.cn-beijing.volces.com/api/v3 | doubao-seed-2-0-prodeepseek-v4-proqwen3.6-plus |
| Baidu Qianfan | OpenAI Compatible | https://qianfan.baidubce.com/v2 | ernie-5.0deepseek-v4-proqwen3.6-plus |
| SiliconFlow | OpenAI Compatible | https://api.siliconflow.cn/v1 | Qwen/Qwen2-72B-Instructdeepseek-ai/DeepSeek-V2-ChatTHUDM/glm-4-9b-chat |
Context length, pricing tiers, and modality support are subject to each platform's latest documentation and console; listed prices are ¥/million tokens, cache-hit prices may be lower.