How to add 65 ChineseModels text models to your coding agents

ChineseModels currently publicly lists 65 text/chat models, out of 79 total models. This article's headline, description and model directory are generated from the same public catalog as Model Square. Image and video generators are excluded from the text count.

The count follows catalog refreshes, including visibility changes; it is not frozen to the number available on publication day. The public catalog normally refreshes upstream pricing data on a 30-minute cache and can use a fallback snapshot during an outage. Your authenticated API key may have different access.

Pick your coding client

• Codex: install the Windows adapter, then launch a separate ChineseModels profile. Your normal OpenAI picker, including Astra, stays intact.
• Kimi Code: add native OpenAI-compatible model aliases alongside your current Kimi login.
• DeepSeek Harness: configure a custom model API provider and select text entries from discovery.
• Claude Code: an experimental Anthropic-format gateway guide. Anthropic does not support using non-Claude models through Claude Code.
• Hermes Agent: merge a named provider with an explicit, filtered text-model list.
• OpenClaw: merge an OpenAI Completions provider and register its model IDs without replacing existing agent defaults.

These are API connections, not downloads of model weights. A public model listing does not prove every coding-client feature works with that model.

Add all available text models

Create a ChineseModels key in your account, then open your client's guide and copy its setup prompt. Let the assistant inspect your installed version, back up configuration and obtain the key through a local secure prompt or secret store.

The account-level source of truth is authenticated GET /v1/models. Import exact model IDs and filter image/video generators before creating picker entries. Do not hard-code the number from this page into an installer.

The Codex and Kimi Windows installers already fetch that authenticated list. For the other clients, use the documented provider configuration and merge the filtered IDs. Where a client supports only one custom picker entry or needs a separate lineup schema, follow that client's guide instead of assuming all pickers work alike.

Preserve existing models and credentials

Keep Codex's ChineseModels profile separate from the normal OpenAI provider. For Kimi, preserve the managed login and default model. In the other clients, merge a named provider or use a dedicated launcher and keep unrelated settings unchanged.

Never publish a key in a prompt, repository, screenshot or URL. The examples reference an environment variable or the client's credential store.

Test the capabilities you need

Start with a short reply. Then test streamed output and a tool call followed by its result. Only enable thinking options whose wire parameters the client and model both support.

Codex and Kimi have documented live connectivity checks in their guides. The other client guides follow official configuration documentation and still require end-to-end validation for your selected model. Model-specific setup instructions are not a compatibility certification.

Keep the list current

Rerun the Codex model-refresh command or the Kimi installer. For other clients, refresh authenticated discovery and merge the new text list. Restart the client or open a fresh session where required.

The number on this article updates as the public catalog refreshes. A local picker updates only when you refresh its configuration; changing the website count does not silently modify anyone's computer.

Choose one model first

Every link below opens a Codex guide named for that model, with an exact launch command and copyable prompt. Model Square offers equivalent setup links for all six clients.

• deepseek-flash
• DeepSeek-V3.2
• DeepSeek-V4-Flash
• DeepSeek-V4-Flash-0731
• DeepSeek-V4-Flash-Vision-Exp
• DeepSeek-V4-Pro
• DeepSeek-V4-Pro-0813
• doubao-1-5-lite-32k-250115
• doubao-1-5-pro-32k-250115
• doubao-seed-1-6-250615
• doubao-seed-1-6-flash-250828
• doubao-seed-1-8-251228
• Doubao-Seed-1.6
• Doubao-Seed-1.6-flash
• Doubao-Seed-1.6-vision
• doubao-seed-2-0-code-preview-260215
• doubao-seed-2-0-lite-260215
• doubao-seed-2-0-lite-260428
• doubao-seed-2-0-mini-260215
• doubao-seed-2-0-mini-260428
• doubao-seed-2-0-pro-260215
• doubao-seed-2-1-pro-260628
• GLM-5
• GLM-5-Turbo
• GLM-5.1
• GLM-5.2
• GLM-5.3
• glm-5.3-flash
• GLM-5V-Turbo
• Hy3
• Hy3-preview
• Kimi-K2.5
• Kimi-K2.6
• Kimi-K2.7-Code
• Kimi-K3
• MiniMax-M2.5
• MiniMax-M2.7
• MiniMax-M3
• qwen-flash
• qwen-long
• qwen-mt-turbo
• qwen-plus
• qwen-plus-latest
• qwen3-coder-plus
• qwen3-max
• qwen3-max-preview
• qwen3-omni-flash
• qwen3-vl-flash
• qwen3-vl-plus
• Qwen3.5-122B-A10B
• Qwen3.5-27B
• Qwen3.5-35B-A3B
• Qwen3.5-397B-A17B
• qwen3.5-flash
• qwen3.5-plus
• Qwen3.6-27B
• Qwen3.6-35B-A3B
• qwen3.6-flash
• qwen3.6-max-preview
• qwen3.6-plus
• qwen3.7-flash
• qwen3.7-max
• qwen3.7-plus
• qwen3.8-flash
• qwen3.8-max

Documentation and operators

The documentation hub gathers the client guides, API reference, thinking limitations and model refresh steps. Operators can manage the Dolphin credential in the authenticated upstream provider settings. Adding that control does not rotate the existing key.