Anthropic

Native Protocol Support

All Anthropic models on this site can be accessed via the native /v1/messages protocol. For example, to activate the 1M-token context, you can use the following request:

curl https://api.highwayapi.ai/anthropic/v1/messages \
  -H "x-api-key: $API_KEY" \
  -H "anthropic-beta: context-1m-2025-08-07" \
  -H "content-type: application/json" \
  -d '{
    "model": "claude-sonnet-4-20250514",
    "max_tokens": 1024,
    "messages": [
      {"role": "user", "content": "Process this large document..."}
    ]
  }'

Prompt caching

This site supports using Prompt caching via either the Anthropic protocol or the OpenAI-compatible protocol. For details, see the documentation Model Features - Prompt caching.

Extended thinking

Currently, the thinking process can only be controlled via the Anthropic protocol.

curl https://api.highwayapi.ai/anthropic/v1/messages \
     -H "x-api-key: $API_KEY" \
     -H "content-type: application/json" \
     -d \
'{
    "model": "claude-sonnet-4-20250514",
    "max_tokens": 16000,
    "thinking": {
        "type": "enabled",
        "budget_tokens": 10000
    },
    "messages": [
        {
            "role": "user",
            "content": "Are there an infinite number of prime numbers such that n mod 4 == 3?"
        }
    ]
}'

Tools

Currently, only Bash and Text editor are supported. Computer use, Web fetch, Web search, and others are not supported yet. For usage, refer to the official documentation.

Using Claude Code

See Third-Party Tool Configuration - Claude Code.

Last modified on June 10, 2026