# Model Gallery

> The full list of models available on the LMU AI API (Chinese LLMs / Claude / OpenAI / Gemini / Grok). Click a model name to copy its ID and paste it into your tool config.

URL: https://docs.lmuai.ai/docs/guide/models



Below are all the model names the LMU AI API currently supports. **Click any card to copy its model ID**, then paste it straight into `settings.json`, `config.toml`, or a tool's model selector.

<Callout type="info" title="How to use a model name">
  Paste the copied model ID into your tool's config:

  * **Claude Code** → the `"model"` field in `settings.json`
  * **Codex CLI** → `model = "..."` in `~/.codex/config.toml`
  * **Cherry Studio / VS Code extension** → paste into the model dropdown
</Callout>

***

## Pick a model by use case [#pick-a-model-by-use-case]

Not sure which model to use? Locate it in the table below, then copy the model ID from the matching section.

| Use case                                                              | Recommended models                                            | Why                                                                                                                 |
| --------------------------------------------------------------------- | ------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------- |
| Complex refactors, large cross-file changes                           | `claude-opus-5` · `gpt-6-astra`                               | Each vendor's flagship tier, outstanding at code and long text, the default choice for Claude Code / Codex          |
| Everyday coding, bug fixing                                           | `claude-sonnet-5` · `glm-5.3`                                 | Balanced tier, a trade-off between capability and speed, ideal for frequent daily use                               |
| Frequent lightweight tasks (completion, formatting, bulk small edits) | `claude-haiku-4-5` · `deepseek-v4-flash` · `qwen3.8-flash`    | High-speed tier, for simple single tasks called many times                                                          |
| Codex CLI / Codex App toolchain                                       | `gpt-5.6-sol` · `gpt-5.5`                                     | The flagship on the OpenAI protocol side and the default in the Codex CLI config template                           |
| Chinese writing, documentation, conversation                          | `qwen3.8-max` · `glm-5.3` · `kimi-k3`                         | Chinese LLMs with excellent Chinese comprehension at a lower price                                                  |
| Coding in Chinese contexts                                            | `qwen3.7-plus` · `deepseek-v4-pro`                            | Code-oriented Chinese models, callable directly with the account balance                                            |
| Multimodal (mixed image + text input)                                 | `gemini-3.1-pro-preview` · `claude-fable-5-1` · `gpt-6-astra` | Natively multimodal models that accept mixed image + text input; for pure image generation see the image APIs below |

<Callout type="warn" title="Confirm two things before choosing">
  * **Model availability depends on the group**: which models you can actually call is whatever the **Available Models** page in the console shows. For example, the GPT series requires a GPT subscription, or selecting a GPT group under pay-as-you-go billing. &#x2A;*This is independent of which protocol you call with.**
  * **The Base URL must match the protocol**: the protocol is decided by your client and SDK — the Anthropic protocol Base URL is **without** `/v1`, the OpenAI protocol **must include** `/v1`, and Gemini native uses `/v1beta/models/...`. Getting it wrong gives an outright 404; see [API Protocols](/docs/guide/api-protocols).
</Callout>

***

## Chinese models [#chinese-models]

Excellent Chinese comprehension at a lower price, ideal for everyday coding, documentation, and conversation. Like Claude and GPT, **Chinese models can all be called using the account balance**.

> **Protocol**: defaults to the **Anthropic protocol** · Base URL `https://api.lmuai.ai` (**without** `/v1`). See [API Protocols →](/docs/guide/api-protocols)

### Alibaba · Qwen [#alibaba--qwen]

- `qwen3.8-max` — Qwen 3.8 Max — flagship（New）
- `qwen3.8-flash` — Qwen 3.8 Flash — high-speed（New）
- `qwen3.7-plus` — Qwen 3.7 Plus
- `qwen3.7-max` — Qwen 3.7 Max
- `qwen3.6-plus` — Qwen 3.6 Plus


### DeepSeek [#deepseek]

- `deepseek-v4-pro` — DeepSeek V4 Pro — flagship
- `deepseek-v4-flash` — DeepSeek V4 Flash — high-speed
- `deepseek-v3.2` — DeepSeek V3.2


### MiniMax [#minimax]

- `MiniMax-M3` — MiniMax M3（New）
- `MiniMax-M2.7` — MiniMax M2.7
- `MiniMax-M2.7-highspeed` — MiniMax M2.7 high-speed
- `MiniMax-M2.5` — MiniMax M2.5
- `MiniMax-M2.5-highspeed` — MiniMax M2.5 high-speed


### Zhipu · GLM [#zhipu--glm]

- `glm-5.3` — Zhipu GLM-5.3 — flagship（New）
- `glm-5.3-flash` — Zhipu GLM-5.3 Flash — high-speed（New）
- `glm-5.2` — Zhipu GLM-5.2
- `glm-5.1` — Zhipu GLM-5.1
- `glm-5` — Zhipu GLM-5


### Moonshot · Kimi [#moonshot--kimi]

- `kimi-k3` — Kimi K3 — flagship
- `kimi-k2.7-code` — Kimi K2.7 Code — coding（New）
- `kimi-k2.6` — Kimi K2.6
- `kimi-k2.5` — Kimi K2.5


### Xiaomi · MiMo [#xiaomi--mimo]

- `mimo-v2.5-pro` — Xiaomi MiMo V2.5 Pro — flagship
- `mimo-v2.5` — Xiaomi MiMo V2.5


***

## Claude models [#claude-models]

Anthropic's official models, outstanding at code and long text, the default for Claude Code.

> **Protocol**: the **Anthropic protocol** is recommended · Base URL `https://api.lmuai.ai` (without `/v1`). The OpenAI protocol also works; the backend translates the protocol.

- `claude-opus-5` — Claude Opus 5 — flagship
- `claude-sonnet-5` — Claude Sonnet 5 — balanced
- `claude-haiku-4-5` — Claude Haiku 4.5 — high-speed
- `claude-fable-5-1` — Claude Fable 5.1 — multimodal（New）
- `claude-opus-4-8` — Claude Opus 4.8


***

## OpenAI models [#openai-models]

The OpenAI GPT series, for Codex CLI, Codex App, and any tool that supports the OpenAI protocol.

> **Protocol**: the **OpenAI protocol** · Base URL `https://api.lmuai.ai/v1` (**must include** `/v1`).

- `gpt-6-astra` — GPT-6 Astra — flagship（New）
- `gpt-5.6-sol` — GPT-5.6 Sol — flagship
- `gpt-5.6-luna` — GPT-5.6 Luna
- `gpt-5.6-terra` — GPT-5.6 Terra
- `gpt-5.5` — GPT-5.5


***

## Google · Gemini models [#google--gemini-models]

Google's Gemini series — natively multimodal with an ultra-long context window (up to a million tokens), ideal for mixed image + text input and long-document processing.

> **Protocol**: callable with the **OpenAI protocol** (Base URL `https://api.lmuai.ai/v1`, **with** `/v1`) or the **Gemini native protocol** (`/v1beta/models/...`). See [API Protocols →](/docs/guide/api-protocols)

- `gemini-3.1-pro-preview` — Gemini 3.1 Pro Preview — flagship（New）
- `gemini-3.8-flash` — Gemini 3.8 Flash — high-speed（New）
- `gemini-3.7-flash` — Gemini 3.7 Flash — high-speed
- `gemini-3.6-flash` — Gemini 3.6 Flash — high-speed
- `gemini-3.5-flash-lite` — Gemini 3.5 Flash Lite — lightweight


***

## xAI · Grok models [#xai--grok-models]

xAI's Grok series — strong reasoning and multimodal capability with support for ultra-long context.

> **Protocol**: the **OpenAI protocol** · Base URL `https://api.lmuai.ai/v1` (**must include** `/v1`).

- `grok-4.6` — Grok 4.6 — flagship（New）
- `grok-4.5` — Grok 4.5
- `grok-4.3` — Grok 4.3
- `grok-4.20-reasoning` — Grok 4.20 Reasoning — reasoning
- `grok-code-fast-1` — Grok Code Fast 1 — coding / high-speed


***

## All model IDs (plain text) [#all-model-ids-plain-text]

<Callout type="info" title="For on-site search and bulk copying">
  **Claude series**: `claude-opus-5`, `claude-sonnet-5`, `claude-haiku-4-5`, `claude-fable-5-1`, `claude-opus-4-8`

  **OpenAI series**: `gpt-6-astra`, `gpt-5.6-sol`, `gpt-5.6-luna`, `gpt-5.6-terra`, `gpt-5.5`

  **Google Gemini**: `gemini-3.1-pro-preview`, `gemini-3.8-flash`, `gemini-3.7-flash`, `gemini-3.6-flash`, `gemini-3.5-flash-lite`

  **xAI Grok**: `grok-4.6`, `grok-4.5`, `grok-4.3`, `grok-4.20-reasoning`, `grok-code-fast-1`

  **Qwen**: `qwen3.8-max`, `qwen3.8-flash`, `qwen3.7-plus`, `qwen3.7-max`, `qwen3.6-plus`

  **DeepSeek**: `deepseek-v4-pro`, `deepseek-v4-flash`, `deepseek-v3.2`

  **MiniMax**: `MiniMax-M3`, `MiniMax-M2.7`, `MiniMax-M2.7-highspeed`, `MiniMax-M2.5`, `MiniMax-M2.5-highspeed`

  **Zhipu GLM**: `glm-5.3`, `glm-5.3-flash`, `glm-5.2`, `glm-5.1`, `glm-5`

  **Moonshot Kimi**: `kimi-k3`, `kimi-k2.7-code`, `kimi-k2.6`, `kimi-k2.5`

  **Xiaomi MiMo**: `mimo-v2.5-pro`, `mimo-v2.5`

  For image models see [GPT Image](/docs/api/gpt-image), [Gemini Image](/docs/api/gemini-image), and [Grok Image](/docs/api/grok-image).
</Callout>

***

<Callout type="warn" title="Note">
  * The actual available models are whatever the **Available Models** page in the LMU AI console shows; the range may differ by plan.
  * If you have questions about model selection, see the [FAQ](/docs/guide/faq).
</Callout>