## Overview

ModelStack provides access to 50+ models from leading AI providers through a single OpenAI-compatible API. Switch between any model by changing one parameter.

<Note>
  Versioned model variants are also supported. Any model with a supported prefix
  (`claude-`, `gpt-`, `gemini-`, `deepseek-`, `qwen-`, `glm-`) is accepted.
</Note>

## Anthropic

| Model ID            | Display Name      | Context | Best For                                |
| ------------------- | ----------------- | ------- | --------------------------------------- |
| `claude-opus-4-6`   | Claude Opus 4.6   | 1M      | Agents, complex reasoning, long-context |
| `claude-opus-4-7`   | Claude Opus 4.7   | 1M      | Complex coding, agentic tasks (legacy)  |
| `claude-sonnet-4-6` | Claude Sonnet 4.6 | 1M      | Best value — near-Opus at Sonnet price  |
| `claude-haiku-4-5`  | Claude Haiku 4.5  | 200K    | High-volume, low-latency, sub-agents    |

## OpenAI

| Model ID       | Display Name | Context | Best For                                        |
| -------------- | ------------ | ------- | ----------------------------------------------- |
| `gpt-5.5`      | GPT-5.5      | 1.05M   | Frontier reasoning, complex problem-solving     |
| `gpt-5.4`      | GPT-5.4      | 1.05M   | Extended reasoning, professional knowledge work |
| `gpt-5`        | GPT-5        | 400K    | Strong general performance                      |
| `gpt-4.1-mini` | GPT-4.1 Mini | 1.05M   | Cost-sensitive coding, vision                   |
| `gpt-4o-mini`  | GPT-4o Mini  | 128K    | High-volume, cost-optimized tasks               |

## Google

| Model ID                | Display Name           | Context | Best For                                  |
| ----------------------- | ---------------------- | ------- | ----------------------------------------- |
| `gemini-3.1-pro`        | Gemini 3.1 Pro Preview | 1.05M   | Agentic coding, software engineering      |
| `gemini-3.1-flash-lite` | Gemini 3.1 Flash Lite  | 1.05M   | High-volume, low-latency multimodal       |
| `gemini-3-flash`        | Gemini 3 Flash Preview | 1.05M   | Multimodal, computer use, vibe coding     |
| `gemini-2.5-pro`        | Gemini 2.5 Pro         | 1.05M   | Complex reasoning, math, STEM             |
| `gemini-2.5-flash`      | Gemini 2.5 Flash       | 1.05M   | Price-performance, large-scale processing |
| `gemma-4-31b-it`        | Gemma 4 31B            | 262K    | Open-weight reasoning and coding          |

## DeepSeek

| Model ID            | Display Name      | Context | Best For                                       |
| ------------------- | ----------------- | ------- | ---------------------------------------------- |
| `deepseek-v3.2`     | DeepSeek V3.2     | 131K    | Reasoning, coding, math at low cost            |
| `deepseek-v4-pro`   | DeepSeek V4 Pro   | 1.05M   | Agentic workflows, cost-effective intelligence |
| `deepseek-v4-flash` | DeepSeek V4 Flash | 1.05M   | Fast reasoning at low cost                     |

## MiniMax

| Model ID       | Display Name | Context | Best For                                               |
| -------------- | ------------ | ------- | ------------------------------------------------------ |
| `minimax-m2.7` | MiniMax M2.7 | 205K    | Agentic workflows, live debugging, document generation |
| `minimax-m2.5` | MiniMax M2.5 | 205K    | Coding (SWE-Bench 80.2%), office productivity          |
| `minimax-m2.1` | MiniMax M2.1 | 205K    | Coding, agentic workflows                              |

## ZAI (GLM)

| Model ID  | Display Name | Context | Best For                |
| --------- | ------------ | ------- | ----------------------- |
| `glm-5`   | GLM 5        | 203K    | General chat and vision |
| `glm-4.7` | GLM 4.7      | 203K    | General chat and vision |

## Qwen (Alibaba)

| Model ID            | Display Name      | Context | Best For                                          |
| ------------------- | ----------------- | ------- | ------------------------------------------------- |
| `qwen3-coder-next`  | Qwen3 Coder Next  | 262K    | Coding agents, tool use, failure recovery         |
| `qwen3-coder-flash` | Qwen3 Coder Flash | 1M      | Multi-file codebases, repository-scale operations |

## Model Selection Guide

<CardGroup cols={2}>
  <Card title="Best Quality" icon="emoji_events">
    **`gpt-5.5`** or **`claude-opus-4-6`** — Frontier models for the most
    complex reasoning, agents, and long-context work.
  </Card>
  <Card title="Best Balance" icon="balance">
    **`claude-sonnet-4-6`** or **`gemini-3.1-pro`** — Near-flagship performance
    at significantly lower cost.
  </Card>
  <Card title="Best Speed" icon="bolt">
    **`claude-haiku-4-5`** or **`gemini-3.1-flash-lite`** — Sub-10ms latency for
    high-volume, latency-sensitive tasks.
  </Card>
  <Card title="Best Value" icon="paid">
    **`deepseek-v3.2`** or **`gpt-4o-mini`** — Lowest cost per token for
    budget-conscious applications.
  </Card>
</CardGroup>