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

# Prisma AIRS AI Gateway Models

> Open-source pricing and configuration database for 2,300+ LLMs across 35+ providers

> [!NOTE]
>
> **AI Gateway Models** is a free, open-source pricing database for LLMs. No API key required — just query any model's pricing directly.

> **[Explore Models](https://portkey.ai/models)**
>
> Browse 2,300+ models

> **[Model Rankings](https://portkey.ai/rankings)**
>
> Top models by usage

> **[GitHub](https://github.com/Portkey-AI/models)**
>
> View source & contribute

***

## Quick Start

Get pricing for any model with a single API call:

```bash cURL theme={"system"}
curl https://aigw.portkey.ai/model-configs/pricing/openai/gpt-4o
```

```javascript JavaScript theme={"system"}
const response = await fetch(
  'https://aigw.portkey.ai/model-configs/pricing/anthropic/claude-3-5-sonnet-20241022'
);
const pricing = await response.json();

console.log(pricing.pay_as_you_go.request_token.price);  // Input cost
console.log(pricing.pay_as_you_go.response_token.price); // Output cost
```

```python Python theme={"system"}
import requests

pricing = requests.get(
    'https://aigw.portkey.ai/model-configs/pricing/google/gemini-2.0-flash-001'
).json()

input_cost = pricing['pay_as_you_go']['request_token']['price']
output_cost = pricing['pay_as_you_go']['response_token']['price']
```

> [!TIP]
>
> **Need to discover all models for a provider?** Instead of querying each model individually, you can fetch pricing for all models at once:
>
> ```bash theme={"system"}
> curl https://configs.portkey.ai/pricing/{provider}.json
> ```
>
> [See detailed documentation below ↓](#get-all-models-for-a-provider)

***

## Understanding Pricing Units

> [!WARNING]
>
> **Prices are in cents per token, not dollars.**

| API Response | Per 1K Tokens | Per 1M Tokens |
| - | - | - |
| `0.003` | \$0.03 | \$30 |
| `0.00025` | \$0.0025 | \$2.50 |
| `0.0001` | \$0.001 | \$1.00 |

**Calculate cost in dollars:**

```javascript theme={"system"}
const costInDollars = (tokens * priceFromAPI) / 100;
```

***

## API Reference

### Get Model Pricing

Returns pricing configuration for a specific model.

```
GET https://aigw.portkey.ai/model-configs/pricing/{provider}/{model}
```

#### Path Parameters

**provider** `string` *required*

Provider identifier. Use lowercase with hyphens.

Examples: `openai`, `anthropic`, `google`, `azure-openai`, `bedrock`, `together-ai`, `groq`, `deepseek`, `x-ai`, `mistral-ai`, `cohere`, `fireworks-ai`, `perplexity-ai`, `anyscale`, `deepinfra`, `cerebras`

**model** `string` *required*

Model identifier. Use the exact model name as specified by the provider.

Examples: `gpt-4o`, `gpt-4-turbo`, `claude-3-5-sonnet-20241022`, `claude-3-opus-20240229`, `gemini-2.0-flash-001`, `gemini-1.5-pro`

#### Response Schema

**pay\_as\_you\_go** `object`

Token-based pricing. All prices in USD **cents per token**.

**properties**

**request\_token** `object`

Input token pricing.

**properties**

**price** `number`

Price in USD cents per token. Example: `0.00025` = \$0.0025 per 1K tokens

**response\_token** `object`

Output token pricing.

**properties**

**price** `number`

Price in USD cents per token.

**cache\_read\_input\_token** `object`

Prompt cache read pricing (when available).

**properties**

**price** `number`

Price in USD cents per token.

**cache\_write\_input\_token** `object`

Prompt cache write pricing (when available).

**properties**

**price** `number`

Price in USD cents per token.

**request\_audio\_token** `object`

Audio input token pricing (for multimodal models).

**properties**

**price** `number`

Price in USD cents per token.

**response\_audio\_token** `object`

Audio output token pricing (for multimodal models).

**properties**

**price** `number`

Price in USD cents per token.

**additional\_units** `object`

Provider-specific pricing for features beyond tokens.

Common units:

* `web_search` — Web search tool usage
* `file_search` — File search tool usage
* `thinking_token` — Chain-of-thought reasoning tokens
* `image_token` — Image generation/processing
* `video_duration_seconds_*` — Video generation (by resolution)

**image** `object`

Image generation pricing by quality and size (for image models).

Structure: `{ "quality": { "resolution": { "price": number } } }`

**calculate** `object`

Cost calculation formulas for complex pricing scenarios.

**properties**

**request** `object`

Formula for calculating request (input) cost.

**response** `object`

Formula for calculating response (output) cost.

**currency** `string`

Currency code. Always `USD`.

**finetune\_config** `object`

Fine-tuning pricing (when available).

**properties**

**pay\_per\_token** `object`

Training cost per token.

**pay\_per\_hour** `object`

Training cost per hour.

#### Example Responses

```json OpenAI GPT-4o theme={"system"}
{
  "pay_as_you_go": {
    "request_token": { "price": 0.00025 },
    "response_token": { "price": 0.001 },
    "cache_write_input_token": { "price": 0 },
    "cache_read_input_token": { "price": 0.000125 },
    "additional_units": {
      "web_search": { "price": 1 },
      "file_search": { "price": 0.25 }
    }
  },
  "calculate": {
    "request": {
      "operation": "sum",
      "operands": [
        { "operation": "multiply", "operands": [{ "value": "input_tokens" }, { "value": "rates.request_token" }] },
        { "operation": "multiply", "operands": [{ "value": "cache_write_tokens" }, { "value": "rates.cache_write_input_token" }] },
        { "operation": "multiply", "operands": [{ "value": "cache_read_tokens" }, { "value": "rates.cache_read_input_token" }] }
      ]
    },
    "response": {
      "operation": "multiply",
      "operands": [{ "value": "output_tokens" }, { "value": "rates.response_token" }]
    }
  },
  "currency": "USD"
}
```

```json Anthropic Claude 3.5 Sonnet theme={"system"}
{
  "pay_as_you_go": {
    "request_token": { "price": 0.0003 },
    "response_token": { "price": 0.0015 },
    "cache_read_input_token": { "price": 0.00003 },
    "cache_write_input_token": { "price": 0.000375 }
  },
  "currency": "USD"
}
```

```json Google Gemini 2.5 Pro theme={"system"}
{
  "pay_as_you_go": {
    "request_token": { "price": 0.000125 },
    "response_token": { "price": 0.001 },
    "additional_units": {
      "thinking_token": { "price": 0.001 },
      "web_search": { "price": 3.5 },
      "search": { "price": 3.5 }
    }
  },
  "currency": "USD"
}
```

```json OpenAI DALL-E 3 (Image Model) theme={"system"}
{
  "pay_as_you_go": {
    "image": {
      "standard": {
        "1024x1024": { "price": 4 },
        "1024x1792": { "price": 8 },
        "1792x1024": { "price": 8 }
      },
      "hd": {
        "1024x1024": { "price": 8 },
        "1024x1792": { "price": 12 },
        "1792x1024": { "price": 12 }
      }
    }
  },
  "currency": "USD"
}
```

***

### Get Model Configuration

Returns general configuration and capabilities for a specific model.

```
GET https://aigw.portkey.ai/model-configs/general/{provider}/{model}
```

#### Path Parameters

**provider** `string` *required*

Provider identifier (e.g., `openai`, `anthropic`, `google`)

**model** `string` *required*

Model identifier (e.g., `gpt-4o`, `claude-3-opus`)

#### Response Schema

**params** `array`

Model parameters with their constraints.

**item properties**

**key** `string`

Parameter name (e.g., `max_tokens`, `temperature`, `top_p`)

**defaultValue** `any`

Default value for the parameter.

**minValue** `number`

Minimum allowed value.

**maxValue** `number`

Maximum allowed value.

**type** `string`

Parameter type (e.g., `boolean`, `string`, `array-of-strings`)

**options** `array`

Available options for enum-type parameters.

**type** `object`

Model type classification.

**properties**

**primary** `string`

Primary model type: `chat`, `text`, `embedding`, `image`, `audio`, `moderation`

**supported** `array`

Supported features: `tools`, `image`, `cache_control`, `json_mode`

**messages** `object`

Message configuration.

**properties**

**options** `array`

Supported message roles: `system`, `user`, `assistant`, `developer`

**disablePlayground** `boolean`

If `true`, model is not available in the AI Gateway playground.

**isDefault** `boolean`

If `true`, this is the default model for the provider.

#### Example Responses

```json OpenAI GPT-4o theme={"system"}
{
  "params": [
    {
      "key": "max_tokens",
      "maxValue": 16384
    },
    {
      "key": "response_format",
      "defaultValue": null,
      "options": [
        { "value": null, "name": "Text" },
        { "value": "json_object", "name": "JSON Object" },
        { "value": "json_schema", "name": "JSON Schema" }
      ],
      "type": "string"
    },
    {
      "key": "temperature",
      "minValue": 0,
      "maxValue": 2,
      "defaultValue": 1
    }
  ],
  "type": {
    "primary": "chat",
    "supported": ["tools", "image"]
  }
}
```

```json Anthropic Claude 3 Opus theme={"system"}
{
  "params": [
    {
      "key": "max_tokens",
      "maxValue": 4096
    }
  ],
  "type": {
    "primary": "chat",
    "supported": ["tools", "image", "cache_control"]
  },
  "messages": {
    "options": ["system", "user", "assistant"]
  }
}
```

***

### Get All Models for a Provider

Returns pricing configuration for all models from a specific provider in a single response.

```
GET https://configs.portkey.ai/pricing/{provider}.json
```

> [!NOTE]
>
> Use this endpoint to discover all available models and their pricing for a provider, instead of querying each model individually.

#### Path Parameters

**provider** `string` *required*

Provider identifier. Use lowercase with hyphens.

Examples: `openai`, `anthropic`, `google`, `bedrock`, `azure-openai`, `together-ai`, `groq`, `deepseek`, `x-ai`, `mistral-ai`

#### Response Schema

Returns a JSON object where:

* Keys are model identifiers
* Values contain `pricing_config` objects with the same structure as the individual model pricing endpoint
* A `default` key provides the base pricing template

**\{model-id}** `object`

Pricing configuration for a specific model.

**properties**

**pricing\_config** `object`

Contains the same pricing structure as the individual model endpoint: `pay_as_you_go`, `calculate`, `currency`, etc.

#### Example Response

```bash cURL theme={"system"}
curl --request GET \
  --url https://configs.portkey.ai/pricing/bedrock.json
```

```javascript JavaScript theme={"system"}
const response = await fetch('https://configs.portkey.ai/pricing/bedrock.json');
const allModels = await response.json();

// Access specific model pricing
const claudePrice = allModels['anthropic.claude-3-5-sonnet-20241022-v2:0'];
console.log(claudePrice.pricing_config.pay_as_you_go);
```

```python Python theme={"system"}
import requests

all_models = requests.get(
    'https://configs.portkey.ai/pricing/bedrock.json'
).json()

# List all available models
for model_id in all_models.keys():
    if model_id != 'default':
        print(model_id)
```

> **Sample Response (Bedrock)**
>
> ```json theme={"system"}
> {
>   "default": {
>     "pricing_config": {
>       "pay_as_you_go": {
>         "request_token": { "price": 0 },
>         "response_token": { "price": 0 }
>       },
>       "calculate": {
>         "request": {
>           "operation": "multiply",
>           "operands": [
>             { "value": "input_tokens" },
>             { "value": "rates.request_token" }
>           ]
>         },
>         "response": {
>           "operation": "multiply",
>           "operands": [
>             { "value": "output_tokens" },
>             { "value": "rates.response_token" }
>           ]
>         }
>       },
>       "currency": "USD"
>     }
>   },
>   "anthropic.claude-3-5-sonnet-20241022-v2:0": {
>     "pricing_config": {
>       "pay_as_you_go": {
>         "request_token": { "price": 0.0003 },
>         "response_token": { "price": 0.0015 },
>         "cache_write_input_token": { "price": 0.000375 },
>         "cache_read_input_token": { "price": 0.00003 }
>       }
>     }
>   },
>   "anthropic.claude-3-haiku-20240307-v1:0": {
>     "pricing_config": {
>       "pay_as_you_go": {
>         "request_token": { "price": 0.000025 },
>         "response_token": { "price": 0.000125 }
>       }
>     }
>   },
>   "meta.llama3-1-405b-instruct-v1:0": {
>     "pricing_config": {
>       "pay_as_you_go": {
>         "request_token": { "price": 0.000532 },
>         "response_token": { "price": 0.0016 }
>       }
>     }
>   },
>   "amazon.nova-pro-v1:0": {
>     "pricing_config": {
>       "pay_as_you_go": {
>         "request_token": { "price": 0.00008 },
>         "response_token": { "price": 0.00002 },
>         "cache_read_input_token": { "price": 0.00004 },
>         "cache_write_input_token": { "price": 0.00016 }
>       }
>     }
>   }
> }
> ```

***

## Additional Units Reference

Provider-specific pricing for features beyond standard token costs:

| Unit | Description | Providers | Price Range (¢) |
| - | - | - | - |
| `web_search` | Web search tool usage | OpenAI, Azure, Google, Vertex AI, Perplexity | 0.5 - 3.5 |
| `file_search` | File search tool usage | OpenAI, Azure | 0.25 |
| `search` | Google search grounding | Google, Vertex AI | 1.4 - 3.5 |
| `thinking_token` | Chain-of-thought reasoning | Google, Vertex AI | 0.00004 - 0.0012 |
| `image_token` | Image processing tokens | Google, Vertex AI | 0.003 |
| `image_1k` | Image generation (1K units) | Google | 3.9 |
| `megapixels` | Image generation by megapixel | Together AI | 0.0027 - 0.08 |
| `video_seconds` | Video generation | Vertex AI | 10 - 50 |
| `video_duration_seconds_720_1280` | Video (720p) | OpenAI Sora | 10 - 30 |
| `video_duration_seconds_1080_1920` | Video (1080p) | OpenAI Sora | 50 |
| `routing_units` | Azure OpenAI routing | Azure OpenAI | 0.000014 |
| `input_image` | Image input | Vertex AI | 0.01 |
| `input_video_essential` | Video input (essential) | Vertex AI | 0.05 |
| `input_video_standard` | Video input (standard) | Vertex AI | 0.1 |
| `input_video_plus` | Video input (plus) | Vertex AI | 0.2 |

> **Perplexity Context-Based Pricing**
>
> Perplexity has context-dependent web search pricing:
>
> | Unit | Price (¢) |
> | - | - |
> | `web_search_low_context` | 0.5 - 0.6 |
> | `web_search_medium_context` | 0.8 - 1.0 |
> | `web_search_high_context` | 1.2 - 1.4 |

***

## Supported Providers

> **40+ providers supported**
>
> AI21, Anthropic, Anyscale, Azure AI, Azure OpenAI, AWS Bedrock, Cerebras, Cohere, Dashscope, Deepbricks, DeepInfra, DeepSeek, Fireworks AI, GitHub, Google, Groq, Inference.net, Jina, Lambda, Lemonfox AI, Mistral AI, MonsterAPI, Nebius, Nomic, Novita AI, OpenAI, OpenRouter, Oracle, PaLM, Perplexity AI, Predibase, Reka AI, Sagemaker, Segmind, Stability AI, Together AI, TypeSafe, Vertex AI, Workers AI, X.AI, Zhipu

***

## Use with AI Gateway

The models database powers automatic cost tracking for all requests through the AI Gateway. When you make requests via the AI Gateway, costs are calculated automatically using this pricing data.

> **[Cost Analytics](https://docs.portkey.ai/docs/aigw/product/observability/analytics.md)**
>
> View real-time spend across all your LLM usage

> **[Budget Limits](https://docs.portkey.ai/docs/aigw/product/model-catalog/integrations.md#3-budget--rate-limits)**
>
> Set spending caps that automatically enforce limits

If you have negotiated rates with a provider, you can override the default pricing:

> **[Custom Pricing](https://docs.portkey.ai/docs/aigw/product/model-catalog/model-overrides.md)**
>
> Set custom input/output costs to match your contracts


## Related topics

- [Security @ Prisma AIRS AI Gateway](https://docs.portkey.ai/docs/aigw/product/enterprise-offering/security.md)
- [Palo Alto Networks Prisma AIRS](https://docs.portkey.ai/docs/aigw/integrations/guardrails/palo-alto-panw-prisma.md)
- [Prisma AIRS AI Gateway: Major Performance Improvements Incoming](https://docs.portkey.ai/docs/aigw/self-hosting/major-performance-improvements.md)
- [GitHub Models](https://docs.portkey.ai/docs/aigw/integrations/llms/github.md)
- [Adding Custom Models](https://docs.portkey.ai/docs/aigw/product/model-catalog/custom-models.md)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.