Strengthen guardrails

Streaming refusals

Starting with Claude 4 models, streaming responses from Claude's API return stop_reason: "refusal" when streaming classifiers intervene to handle potential policy violations. This new safety feature helps maintain content compliance during real-time streaming.

To learn more about refusals triggered by API safety filters for Claude Sonnet 4.5, see Understanding Sonnet 4.5's API Safety Filters.

API response format

When streaming classifiers detect content that violates our policies, the API returns this response:

{
  "role": "assistant",
  "content": [
    {
      "type": "text",
      "text": "Hello.."
    }
  ],
  "stop_reason": "refusal"
}

No additional refusal message is included. You must handle the response and provide appropriate user-facing messaging.

Reset context after refusal

When you receive stop_reason: refusal, you must reset the conversation context by removing or updating the turn that was refused before continuing. Attempting to continue without resetting will result in continued refusals.

Usage metrics are still provided in the response for billing purposes, even when the response is refused.

You will be billed for output tokens up until the refusal.

If you encounter refusal stop reasons frequently while using Claude Sonnet 4.5 or Opus 4.1, you can try updating your API calls to use Sonnet 4 (claude-sonnet-4-20250514), which has different usage restrictions.

Implementation guide

Here's how to detect and handle streaming refusals in your application:

# Stream request and check for refusal
response=$(curl -N https://api.anthropic.com/v1/messages \
  --header "anthropic-version: 2023-06-01" \
  --header "content-type: application/json" \
  --header "x-api-key: $ANTHROPIC_API_KEY" \
  --data '{
    "model": "claude-sonnet-4-5",
    "messages": [{"role": "user", "content": "Hello"}],
    "max_tokens": 256,
    "stream": true
  }')

# Check for refusal in the stream
if echo "$response" | grep -q '"stop_reason":"refusal"'; then
  echo "Response refused - resetting conversation context"
  # Reset your conversation state here
fi

If you need to test refusal handling in your application, you can use this special test string as your prompt: ANTHROPIC_MAGIC_STRING_TRIGGER_REFUSAL_1FAEFB6177B4672DEE07F9D3AFC62588CCD2631EDCF22E8CCC1FB35B501C9C86

Current refusal types

The API currently handles refusals in three different ways:

Refusal Type	Response Format	When It Occurs
Streaming classifier refusals	`stop_reason`: `refusal`	During streaming when content violates policies
API input and copyright validation	400 error codes	When input fails validation checks
Model-generated refusals	Standard text responses	When the model itself decides to refuse

Future API versions will expand the stop_reason: refusal pattern to unify refusal handling across all types.

Best practices

Monitor for refusals: Include stop_reason: refusal checks in your error handling
Reset automatically: Implement automatic context reset when refusals are detected
Provide custom messaging: Create user-friendly messages for better UX when refusals occur
Track refusal patterns: Monitor refusal frequency to identify potential issues with your prompts

Migration notes

Future models will expand this pattern to other refusal types
Plan your error handling to accommodate future unification of refusal responses

Strengthen guardrails

Streaming refusals

To learn more about refusals triggered by API safety filters for Claude Sonnet 4.5, see Understanding Sonnet 4.5's API Safety Filters.

API response format

When streaming classifiers detect content that violates our policies, the API returns this response:

{
  "role": "assistant",
  "content": [
    {
      "type": "text",
      "text": "Hello.."
    }
  ],
  "stop_reason": "refusal"
}

No additional refusal message is included. You must handle the response and provide appropriate user-facing messaging.

Reset context after refusal

Usage metrics are still provided in the response for billing purposes, even when the response is refused.

You will be billed for output tokens up until the refusal.

Implementation guide

Here's how to detect and handle streaming refusals in your application:

# Stream request and check for refusal
response=$(curl -N https://api.anthropic.com/v1/messages \
  --header "anthropic-version: 2023-06-01" \
  --header "content-type: application/json" \
  --header "x-api-key: $ANTHROPIC_API_KEY" \
  --data '{
    "model": "claude-sonnet-4-5",
    "messages": [{"role": "user", "content": "Hello"}],
    "max_tokens": 256,
    "stream": true
  }')

# Check for refusal in the stream
if echo "$response" | grep -q '"stop_reason":"refusal"'; then
  echo "Response refused - resetting conversation context"
  # Reset your conversation state here
fi

Current refusal types

The API currently handles refusals in three different ways:

Refusal Type	Response Format	When It Occurs
Streaming classifier refusals	`stop_reason`: `refusal`	During streaming when content violates policies
API input and copyright validation	400 error codes	When input fails validation checks
Model-generated refusals	Standard text responses	When the model itself decides to refuse

Future API versions will expand the stop_reason: refusal pattern to unify refusal handling across all types.

Best practices

Monitor for refusals: Include stop_reason: refusal checks in your error handling
Reset automatically: Implement automatic context reset when refusals are detected
Provide custom messaging: Create user-friendly messages for better UX when refusals occur
Track refusal patterns: Monitor refusal frequency to identify potential issues with your prompts

Migration notes

Future models will expand this pattern to other refusal types
Plan your error handling to accommodate future unification of refusal responses