> ## Documentation Index
> Fetch the complete documentation index at: https://docs.gravitex.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Supported Models

> GravitexAI model catalog — 100+ frontier models from Anthropic, OpenAI, Google, Qwen, BytePlus, DeepSeek, GLM, MiniMax, MoonShot with a single OpenAI-compatible API

## One API. 100+ Models.

**GravitexAI** is the unified AI model platform built for agents. A single **OpenAI SDK–compatible** interface routes you to 100+ frontier models — no per-provider plumbing, no subscriptions, pay-as-you-go.

<Note>
  **Better prices, higher uptime, no subscriptions.** Distributed infrastructure with automatic failover routes around outages, while organization-level data policies give you fine-grained control over which providers can see your data.
</Note>

## ⭐ Latest Releases

| Model                | Provider  | Best for                                           |
| -------------------- | --------- | -------------------------------------------------- |
| **Claude Fable 5** ⭐ | Anthropic | Complex reasoning, long-horizon agents, 1M context |
| **Qwen3.7-Plus** ⭐   | Alibaba   | Multimodal agents, GUI automation, visual coding   |
| **Claude Opus 4.8**  | Anthropic | Coding, agent workflows, professional tasks        |

For the complete catalog and live pricing, visit the [GravitexAI model catalog](https://maas.gravitex.ai/#/api-models).

## Trusted by the World's Leading Providers

GravitexAI integrates with the top model providers — all behind the same OpenAI-compatible API:

> **Anthropic** · **OpenAI** · **Google** · **Qwen** · **BytePlus (Doubao)** · **DeepSeek** · **GLM** · **MiniMax** · **MoonShot (Kimi)**

***

## Model Catalog

### 🎭 Anthropic Claude

| Model                 | Model ID                   | Context | Highlights                                                                                   | Recommended For                            |
| --------------------- | -------------------------- | ------- | -------------------------------------------------------------------------------------------- | ------------------------------------------ |
| **Claude Fable 5** ⭐  | claude-fable-5             | 1M      | Mythos-class flagship for complex reasoning and long-horizon agent tasks                     | Complex reasoning, long-horizon agents     |
| **Claude Opus 4.8**   | claude-opus-4-8            | 1M      | Upgraded Opus line with stronger coding and agent reliability                                | Coding, agents, professional workflows     |
| **Claude Opus 4.7**   | claude-opus-4-7            | 1M      | Flagship for complex reasoning and agentic coding                                            | Complex reasoning, production code, agents |
| **Claude Opus 4.6**   | claude-opus-4-6            | 1M      | Anthropic's flagship intelligence model; sets the bar for production code and complex agents | Conversation, reasoning, coding, writing   |
| **Claude Sonnet 4.6** | claude-sonnet-4-6          | 1M      | Latest Sonnet mainline with strong all-round capabilities                                    | Conversation, thinking, writing, coding    |
| **Claude Sonnet 4.5** | claude-sonnet-4-5-20250929 | 200K    | Improved tool use, memory and context handling                                               | Conversation, thinking, writing, coding    |
| **Claude Haiku 4.5**  | claude-haiku-4-5-20251001  | 200K    | Anthropic's most efficient model — near-frontier performance, fast and cheap                 | High-throughput chat, light tasks          |
| **Claude Opus 4.5**   | claude-opus-4-5-20251101   | 200K    | Previous flagship, stable and reliable                                                       | Conversation, thinking, writing, coding    |

### 🤖 OpenAI

#### GPT Series

| Model            | Model ID     | Context | Highlights                                                              | Recommended For                         |
| ---------------- | ------------ | ------- | ----------------------------------------------------------------------- | --------------------------------------- |
| **GPT-5.5** ⭐    | gpt-5.5      | 1M      | OpenAI's latest mainline with stronger reasoning and overall capability | Conversation, thinking, coding, writing |
| **GPT 5.4**      | gpt-5.4      | 1M      | New-generation mainline for complex professional tasks                  | Conversation, thinking, coding, writing |
| **GPT 5.4 Pro**  | gpt-5.4-pro  | 1M      | High-performance variant for difficult reasoning and engineering        | Conversation, thinking, coding          |
| **GPT 5.4 Mini** | gpt-5.4-mini | 1M      | Balanced performance and cost                                           | Conversation, writing, coding           |
| **GPT 5.4 Nano** | gpt-5.4-nano | 1M      | Lightweight low-cost variant for high-throughput                        | Conversation, writing                   |
| **GPT 5.3 Chat** | gpt-5.3-chat | 400K    | Stable general chat model                                               | Conversation, writing                   |

#### Image Generation

| Model             | Model ID    | Highlights                           |
| ----------------- | ----------- | ------------------------------------ |
| **GPT-Image-2** ⭐ | gpt-image-2 | OpenAI's next-generation image model |

### 🌟 Google Gemini

| Model                              | Model ID                       | Context | Highlights                                                     | Recommended For                          |
| ---------------------------------- | ------------------------------ | ------- | -------------------------------------------------------------- | ---------------------------------------- |
| **Gemini 3.5 Flash** ⭐             | gemini-3.5-flash               | 1M      | Latest Gemini Flash — high throughput, low latency             | Tool calls, high-concurrency chat        |
| **Gemini 3.1 Pro Preview**         | gemini-3.1-pro-preview         | 1M      | Google's upgraded core intelligence model for complex tasks    | Conversation, reasoning, writing, coding |
| **Gemini 3.1 Flash Lite Preview**  | gemini-3.1-flash-lite-preview  | 1M      | High-value fast model for high-throughput workloads            | Conversation, writing                    |
| **Gemini 3.1 Flash Image Preview** | gemini-3.1-flash-image-preview | 1M      | Multimodal variant focused on image understanding & generation | Vision, text-to-image, multimodal        |
| **Gemini 3 Pro Image Preview**     | gemini-3-pro-image-preview     | 1M      | Gemini image-focused variant                                   | Text-to-image, image generation          |
| **Gemini 3 Flash Preview**         | gemini-3-flash-preview         | 1M      | Fast response model for interactive and tool workflows         | Conversation, writing, coding            |

### 🐉 Alibaba Qwen

| Model                   | Model ID            | Context | Highlights                                                                  | Recommended For                         |
| ----------------------- | ------------------- | ------- | --------------------------------------------------------------------------- | --------------------------------------- |
| **Qwen3.7-Plus** ⭐      | qwen3.7-plus        | 1M      | Multimodal agent model with text, image, and video input and GUI automation | Multimodal agents, visual coding        |
| **Qwen3.7-Max**         | qwen3.7-max         | 1M      | Qwen3.7 flagship text model for long-horizon autonomous agent workflows     | Conversation, reasoning, coding, agents |
| **Qwen3.6-Max-Preview** | qwen3.6-max-preview | 262K    | Qwen3.6 flagship preview with stronger agentic coding                       | Conversation, reasoning, coding         |
| **Qwen3 Coder Plus**    | qwen3-coder-plus    | 1M      | Qwen3-based coding model with strong agent and tool capabilities            | Coding, tool use                        |
| **Qwen Image Plus**     | qwen-image-plus     | 32K     | Qwen image model with strong text rendering                                 | Text-to-image, image generation         |

### 🌋 BytePlus Doubao Seed

| Model                            | Model ID                            | Context     | Highlights                                              | Recommended For                 |
| -------------------------------- | ----------------------------------- | ----------- | ------------------------------------------------------- | ------------------------------- |
| **Doubao Seed 2.0 Pro** ⭐        | doubao-seed-2-0-pro-260215          | See console | Latest mainline Doubao with stronger overall capability | Conversation, reasoning, coding |
| **Doubao Seed 2.0 Code Preview** | doubao-seed-2-0-code-preview-260215 | See console | Preview model optimized for coding tasks                | Coding, tool use                |
| **Doubao Seed 2.0 Lite**         | doubao-seed-2-0-lite-260215         | See console | Lightweight model for cost-sensitive usage              | Conversation, writing           |
| **Doubao Seed 2.0 Mini**         | doubao-seed-2-0-mini-260215         | See console | Smaller fast model for low-latency scenarios            | Conversation, fast response     |
| **Doubao Seed 1.8**              | doubao-seed-1-8-251228              | See console | Stable general model                                    | Conversation, writing           |

#### Image Generation

| Model                     | Model ID                   | Highlights                                                                      | Recommended For               |
| ------------------------- | -------------------------- | ------------------------------------------------------------------------------- | ----------------------------- |
| **Doubao Seedream 5.0** ⭐ | doubao-seedream-5-0-260128 | Latest BytePlus image model with stronger instruction following and consistency | Text-to-image, image-to-image |

### 🎬 Video Generation

| Model                | Model ID                         | Specs                 | Highlights                                                           | Recommended For                   |
| -------------------- | -------------------------------- | --------------------- | -------------------------------------------------------------------- | --------------------------------- |
| **HappyHorse 1.0** ⭐ | happyhorse-1.0-t2v / i2v / r2v   | 720P/1080P, 3–15s     | Physically realistic, smooth motion; R2V supports up to 9 image refs | Text / image / reference-to-video |
| **Wan2.7**           | wan2.7-t2v / i2v / r2v           | 720P/1080P, up to 15s | Multi-shot storytelling, I2V continuation, multimodal references     | Text / image / reference-to-video |
| **Kling V3**         | kling-v3 / pro / omni / omni-pro | 720P/1080P, 3–15s     | Multi-shot storyboarding, audio sync; Omni adds video references     | Narrative video creation          |
| **Seedance 2.0**     | seedance-2-0 / fast / NSFW       | 480p–1080p, 4–15s     | Multimodal references, asset libraries, face consistency             | Multimodal video generation       |

<Tip>
  See [Submit Video Task](/en/api-reference/endpoint/submit-video-task) and model docs such as [Wan 2.7](/en/api-reference/endpoint/wan2.7) and [HappyHorse](/en/api-reference/endpoint/happyhorse).
</Tip>

### 🔍 DeepSeek

| Model                    | Model ID             | Context | Highlights                                                          | Recommended For                         |
| ------------------------ | -------------------- | ------- | ------------------------------------------------------------------- | --------------------------------------- |
| **DeepSeek V4 Pro** ⭐    | deepseek-v4-pro      | 1M      | Million-token MoE flagship with `-none` / `-max` reasoning suffixes | Agents, complex reasoning, coding       |
| **DeepSeek V4 Flash**    | deepseek-v4-flash    | 1M      | Cost-effective MoE variant for long-context agents and reasoning    | High-throughput, agents, coding         |
| **DeepSeek V3.2 251201** | deepseek-v3-2-251201 | 128K    | Stable hybrid-reasoning model                                       | Conversation, thinking, writing, coding |
| **DeepSeek V3 250324**   | deepseek-v3-250324   | 128K    | Stable and cost-effective general model                             | General use                             |
| **DeepSeek R1 250528**   | deepseek-r1-250528   | 64K     | Reasoning-focused model for logic tasks                             | Math, reasoning                         |

### 💎 Zhipu GLM

| Model         | Model ID | Context     | Highlights                                                                          | Recommended For                                       |
| ------------- | -------- | ----------- | ----------------------------------------------------------------------------------- | ----------------------------------------------------- |
| **glm-5.1** ⭐ | glm-5.1  | 200K        | Agentic engineering flagship with long-horizon coding and SWE-Bench Pro performance | Complex software engineering, long-horizon agents     |
| **glm-5**     | glm-5    | See console | Zhipu mainline chat model                                                           | Conversation, writing, coding                         |
| **glm-4.7**   | glm-4.7  | 200K        | Latest flagship — stronger coding and multi-step reasoning, 355B params             | Conversation, long-horizon planning, coding, tool use |

### ✨ MiniMax

| Model               | Model ID        | Context     | Highlights                                                 | Recommended For                   |
| ------------------- | --------------- | ----------- | ---------------------------------------------------------- | --------------------------------- |
| **MiniMax-M2.5** ⭐  | MiniMax-M2.5    | 205K        | Agent flagship, 80.2% SWE-Bench Verified, tool use and MCP | Coding, agents, complex reasoning |
| **MiniMax Text-01** | minimax-text-01 | See console | Long-context MoE model for complex workloads               | Long text, complex reasoning      |

<Tip>
  The MiniMax full catalog and live pricing are listed on the [GravitexAI model catalog](https://maas.gravitex.ai/#/api-models).
</Tip>

### 🌙 Moonshot Kimi

| Model              | Model ID       | Context     | Highlights                                                               | Recommended For                           |
| ------------------ | -------------- | ----------- | ------------------------------------------------------------------------ | ----------------------------------------- |
| **kimi-k2.6** ⭐    | kimi-k2.6      | See console | Latest Kimi with 4,000+ tool calls and 12-hour-class long-horizon coding | Complex coding, multi-agent collaboration |
| **kimi-k2.5**      | kimi-k2.5      | 256K        | Native multimodal (vision + text), thinking and non-thinking modes       | Conversation, thinking                    |
| **Kimi K2 250905** | kimi-k2-250905 | 256K        | MoE architecture, 1T params (32B active), strong coding & agent ability  | Conversation, thinking, writing, coding   |

## 🎯 Platform Capabilities

<Columns cols={2}>
  <Card title="One API for any model" icon="plug">
    Fully OpenAI SDK–compatible. Swap providers without changing a single line of code.
  </Card>

  <Card title="Higher availability" icon="shield">
    Distributed infrastructure with automatic failover routes around outages.
  </Card>

  <Card title="Price & performance" icon="bolt">
    Edge infrastructure for minimal latency. Transparent token-level pricing, pay only for what you use.
  </Card>

  <Card title="Custom data policies" icon="lock">
    Fine-grained privacy controls — pick which providers may access your data.
  </Card>
</Columns>

## 💰 Pricing

### Billing

* **Pay-as-you-go**: Billed by actual Token usage
* **No minimum**: Use what you pay for; balance never expires
* **Real-time billing**: Deducted immediately after each call

### Price Advantages

* Direct upstream routing with competitive pricing vs. provider-direct
* Contact us for bulk pricing

## 🛠️ Usage Guide

### Model Selection

<Tabs>
  <Tab title="Coding">
    * **Top performance**: Claude Fable 5, Claude Opus 4.8, GPT-5.5, Qwen3.7-Max, MiniMax-M2.5
    * **Cost-effective**: Claude Haiku 4.5, GPT 5.4 Mini, DeepSeek V4 Flash, glm-5.1
    * **Alternative**: Gemini 3.5 Flash, Qwen3 Coder Plus, kimi-k2.6, doubao-seed-2-0-code-preview-260215
  </Tab>

  <Tab title="Writing">
    * **Primary**: GPT-5.5, Claude Fable 5, Claude Opus 4.8, Qwen3.7-Plus
    * **Alternative**: Gemini 3.5 Flash, kimi-k2.6, MiniMax-M2.5, doubao-seed-2-0-pro-260215
  </Tab>

  <Tab title="Fast Response">
    * **Primary**: Claude Haiku 4.5, Gemini 3.5 Flash, Gemini 3.1 Flash Lite Preview
    * **Alternative**: glm-5.1, GPT 5.4 Nano, doubao-seed-2-0-lite-260215, DeepSeek V4 Flash
  </Tab>

  <Tab title="Image Generation">
    * **Recommended**: GPT-Image-2, Doubao Seedream 5.0, Qwen Image Plus, Gemini 3.1 Flash Image Preview
  </Tab>

  <Tab title="Video Generation">
    * **Recommended**: HappyHorse 1.0, Wan2.7, Kling V3, Seedance 2.0
  </Tab>

  <Tab title="Long Text">
    * **Ultra-long context**: Gemini 3.5 Flash (1M), Claude Fable 5 (1M), GPT-5.5 (1M), DeepSeek V4 (1M)
    * **Coding**: Claude 4.x series, kimi-k2.6, Qwen3.7-Max
  </Tab>
</Tabs>

### Cost Optimization

1. **Tiered usage**: Use cheaper models for simple tasks, flagships for complex tasks
2. **Test first**: Iterate with smaller models before scaling up
3. **Batch processing**: Use Mini / Lite variants for bulk tasks
4. **Cache reuse**: Cache repeated query results

## 🔗 Related Resources

<Columns cols={2}>
  <Card title="API Docs" icon="book" href="/en/api-reference/openai-sdk">
    OpenAI SDK usage and detailed API spec
  </Card>

  <Card title="Quick Start" icon="rocket" href="/en/quickstart">
    Get your API key and ship your first request
  </Card>
</Columns>

<Tip>
  The model list is continuously updated. For specific model requests or bulk needs, email [bd@gravitex.ai](mailto:bd@gravitex.ai) or browse the [GravitexAI model catalog](https://maas.gravitex.ai/#/api-models).
</Tip>