← Back to daily report

build-with-claude/claude-on-vertex-ai.md

Changed on 2026-02-05 12:43:36 EST

+6 lines added
-5 lines removed
Visual Diff
# Claude on Vertex AI¶

Anthropic's Claude models are now generally available through [Vertex AI](https://cloud.google.com/vertex-ai).¶

---¶

The Vertex API for accessing Claude is nearly-identical to the [Messages API](/docs/en/api/messages) and supports all of the same options, with two key differences:¶

* In Vertex, `model` is not passed in the request body. Instead, it is specified in the Google Cloud endpoint URL.¶
* In Vertex, `anthropic_version` is passed in the request body (rather than as a header), and must be set to the value `vertex-2023-10-16`.¶

Vertex is also supported by Anthropic's official [client SDKs](/docs/en/api/client-sdks). This guide will walk you through the process of making a request to Claude on Vertex AI in either Python or TypeScript.¶

Note that this guide assumes you have already have a GCP project that is able to use Vertex AI. See [using the Claude 3 models from Anthropic](https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/use-claude) for more information on the setup required, as well as a full walkthrough.¶

## Install an SDK for accessing Vertex AI¶

First, install Anthropic's [client SDK](/docs/en/api/client-sdks) for your language of choice.¶

<CodeGroup>¶
```python Python¶
pip install -U google-cloud-aiplatform "anthropic[vertex]"¶
```¶

```typescript TypeScript¶
npm install @anthropic-ai/vertex-sdk¶
```¶
</CodeGroup>¶

## Accessing Vertex AI¶

### Model Availability¶

Note that Anthropic model availability varies by region. Search for "Claude" in the [Vertex AI Model Garden](https://cloud.google.com/model-garden) or go to [Use Claude 3](https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/use-claude) for the latest information.¶

#### API model IDs¶

| Model | Vertex AI API model ID |¶
| ------------------------------ | ------------------------ |¶
| Claude Opus 4.6 | claude-opus-4-6 |¶
| Claude Sonnet 4.5 | claude-sonnet-4-5@20250929 |¶
| Claude Sonnet 4 | claude-sonnet-4@20250514 |¶
| Claude Sonnet 3.7 <Tooltip tooltipContent="Deprecated as of October 28, 2025.">⚠️</Tooltip> | claude-3-7-sonnet@20250219 |¶
| Claude Opus 4.5 | claude-opus-4-5@20251101 |¶
| Claude Opus 4.1 | claude-opus-4-1@20250805 |¶
| Claude Opus 4 | claude-opus-4@20250514 |¶
| Claude Haiku 4.5 | claude-haiku-4-5@20251001 |¶
| Claude Haiku 3.5 <Tooltip tooltipContent="Deprecated as of December 19, 2025.">⚠️</Tooltip> | claude-3-5-haiku@20241022 |¶
| Claude Haiku 3 | claude-3-haiku@20240307 |¶

### Making requests¶

Before running requests you may need to run `gcloud auth application-default login` to authenticate with GCP.¶

The following examples shows how to generate text from Claude on Vertex AI:¶
<CodeGroup>¶

```python Python¶
from anthropic import AnthropicVertex¶

project_id = "MY_PROJECT_ID"¶
region = "global"¶

client = AnthropicVertex(project_id=project_id, region=region)¶

message = client.messages.create(¶
model="claude-sonnet-4-5@20250929opus-4-6",¶
max_tokens=100,¶
messages=[¶
{¶
"role": "user",¶
"content": "Hey Claude!",¶
}¶
],¶
)¶
print(message)¶
```¶

```typescript TypeScript¶
import { AnthropicVertex } from '@anthropic-ai/vertex-sdk';¶

const projectId = 'MY_PROJECT_ID';¶
const region = 'global';¶

// Goes through the standard `google-auth-library` flow.¶
const client = new AnthropicVertex({¶
projectId,¶
region,¶
});¶

async function main() {¶
const result = await client.messages.create({¶
model: 'claude-sonnet-4-5@20250929opus-4-6',¶
max_tokens: 100,¶
messages: [¶
{¶
role: 'user',¶
content: 'Hey Claude!',¶
},¶
],¶
});¶
console.log(JSON.stringify(result, null, 2));¶
}¶

main();¶
```¶

```bash Shell¶
MODEL_ID=claude-sonnet-4-5@20250929opus-4-6
LOCATION=global¶
PROJECT_ID=MY_PROJECT_ID¶

curl \¶
-X POST \¶
-H "Authorization: Bearer $(gcloud auth print-access-token)" \¶
-H "Content-Type: application/json" \¶
https://$LOCATION-aiplatform.googleapis.com/v1/projects/${PROJECT_ID}/locations/${LOCATION}/publishers/anthropic/models/${MODEL_ID}:streamRawPredict -d \¶
'{¶
"anthropic_version": "vertex-2023-10-16",¶
"messages": [{¶
"role": "user",¶
"content": "Hey Claude!"¶
}],¶
"max_tokens": 100,¶
}'¶
```¶
</CodeGroup>¶

See our [client SDKs](/docs/en/api/client-sdks) and the official [Vertex AI docs](https://cloud.google.com/vertex-ai/docs) for more details.¶

## Activity logging¶

Vertex provides a [request-response logging service](https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/request-response-logging) that allows customers to log the prompts and completions associated with your usage. ¶

Anthropic recommends that you log your activity on at least a 30-day rolling basis in order to understand your activity and investigate any potential misuse. ¶

<Note>¶
Turning on this service does not give Google or Anthropic any access to your content.¶
</Note>¶

## Feature support¶
You can find all the features currently supported on Vertex [here](/docs/en/api/overview).¶

## Global vs regional endpoints¶

Starting with **Claude Sonnet 4.5 and all future models**, Google Vertex AI offers two endpoint types:¶

- **Global endpoints**: Dynamic routing for maximum availability¶
- **Regional endpoints**: Guaranteed data routing through specific geographic regions¶

Regional endpoints include a 10% pricing premium over global endpoints.¶

<Note>¶
This applies to Claude Sonnet 4.5 and future models only. Older models (Claude Sonnet 4, Opus 4, and earlier) maintain their existing pricing structures.¶
</Note>¶

### When to use each option¶

**Global endpoints (recommended):**¶
- Provide maximum availability and uptime¶
- Dynamically route requests to regions with available capacity¶
- No pricing premium¶
- Best for applications where data residency is flexible¶
- Only supports pay-as-you-go traffic (provisioned throughput requires regional endpoints)¶

**Regional endpoints:**¶
- Route traffic through specific geographic regions¶
- Required for data residency and compliance requirements¶
- Support both pay-as-you-go and provisioned throughput¶
- 10% pricing premium reflects infrastructure costs for dedicated regional capacity¶

### Implementation¶

**Using global endpoints (recommended):**¶

Set the `region` parameter to `"global"` when initializing the client:¶

<CodeGroup>¶
```python Python¶
from anthropic import AnthropicVertex¶

project_id = "MY_PROJECT_ID"¶
region = "global"¶

client = AnthropicVertex(project_id=project_id, region=region)¶

message = client.messages.create(¶
model="claude-sonnet-4-5@20250929opus-4-6",¶
max_tokens=100,¶
messages=[¶
{¶
"role": "user",¶
"content": "Hey Claude!",¶
}¶
],¶
)¶
print(message)¶
```¶

```typescript TypeScript¶
import { AnthropicVertex } from '@anthropic-ai/vertex-sdk';¶

const projectId = 'MY_PROJECT_ID';¶
const region = 'global';¶

const client = new AnthropicVertex({¶
projectId,¶
region,¶
});¶

const result = await client.messages.create({¶
model: 'claude-sonnet-4-5@20250929opus-4-6',¶
max_tokens: 100,¶
messages: [¶
{¶
role: 'user',¶
content: 'Hey Claude!',¶
},¶
],¶
});¶
```¶
</CodeGroup>¶

**Using regional endpoints:**¶

Specify a specific region like `"us-east1"` or `"europe-west1"`:¶

<CodeGroup>¶
```python Python¶
from anthropic import AnthropicVertex¶

project_id = "MY_PROJECT_ID"¶
region = "us-east1" # Specify a specific region¶

client = AnthropicVertex(project_id=project_id, region=region)¶

message = client.messages.create(¶
model="claude-sonnet-4-5@20250929opus-4-6",¶
max_tokens=100,¶
messages=[¶
{¶
"role": "user",¶
"content": "Hey Claude!",¶
}¶
],¶
)¶
print(message)¶
```¶

```typescript TypeScript¶
import { AnthropicVertex } from '@anthropic-ai/vertex-sdk';¶

const projectId = 'MY_PROJECT_ID';¶
const region = 'us-east1'; // Specify a specific region¶

const client = new AnthropicVertex({¶
projectId,¶
region,¶
});¶

const result = await client.messages.create({¶
model: 'claude-sonnet-4-5@20250929opus-4-6',¶
max_tokens: 100,¶
messages: [¶
{¶
role: 'user',¶
content: 'Hey Claude!',¶
},¶
],¶
});¶
```¶
</CodeGroup>¶

### Additional resources¶

- **Google Vertex AI pricing:** [cloud.google.com/vertex-ai/generative-ai/pricing](https://cloud.google.com/vertex-ai/generative-ai/pricing)¶
- **Claude models documentation:** [Claude on Vertex AI](https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/claude)¶
- **Google blog post:** [Global endpoint for Claude models](https://cloud.google.com/blog/products/ai-machine-learning/global-endpoint-for-claude-models-generally-available-on-vertex-ai)¶
- **Anthropic pricing details:** [Pricing documentation](/docs/en/about-claude/pricing#third-party-platform-pricing)

Unified Diff

--- a/build-with-claude/claude-on-vertex-ai.md
+++ b/build-with-claude/claude-on-vertex-ai.md
@@ -37,6 +37,7 @@
 
 | Model                          | Vertex AI API model ID |
 | ------------------------------ | ------------------------ |
+| Claude Opus 4.6                  | claude-opus-4-6 |
 | Claude Sonnet 4.5              | claude-sonnet-4-5@20250929 |
 | Claude Sonnet 4                | claude-sonnet-4@20250514 |
 | Claude Sonnet 3.7 <Tooltip tooltipContent="Deprecated as of October 28, 2025.">⚠️</Tooltip> | claude-3-7-sonnet@20250219 |
@@ -63,7 +64,7 @@
   client = AnthropicVertex(project_id=project_id, region=region)
 
   message = client.messages.create(
-      model="claude-sonnet-4-5@20250929",
+      model="claude-opus-4-6",
       max_tokens=100,
       messages=[
           {
@@ -89,7 +90,7 @@
 
   async function main() {
     const result = await client.messages.create({
-      model: 'claude-sonnet-4-5@20250929',
+      model: 'claude-opus-4-6',
       max_tokens: 100,
       messages: [
         {
@@ -105,7 +106,7 @@
   ```
 
   ```bash Shell
-  MODEL_ID=claude-sonnet-4-5@20250929
+  MODEL_ID=claude-opus-4-6
   LOCATION=global
   PROJECT_ID=MY_PROJECT_ID
 
@@ -184,7 +185,7 @@
 client = AnthropicVertex(project_id=project_id, region=region)
 
 message = client.messages.create(
-    model="claude-sonnet-4-5@20250929",
+    model="claude-opus-4-6",
     max_tokens=100,
     messages=[
         {
@@ -208,7 +209,7 @@
 });
 
 const result = await client.messages.create({
-  model: 'claude-sonnet-4-5@20250929',
+  model: 'claude-opus-4-6',
   max_tokens: 100,
   messages: [
     {
@@ -234,7 +235,7 @@
 client = AnthropicVertex(project_id=project_id, region=region)
 
 message = client.messages.create(
-    model="claude-sonnet-4-5@20250929",
+    model="claude-opus-4-6",
     max_tokens=100,
     messages=[
         {
@@ -258,7 +259,7 @@
 });
 
 const result = await client.messages.create({
-  model: 'claude-sonnet-4-5@20250929',
+  model: 'claude-opus-4-6',
   max_tokens: 100,
   messages: [
     {