Using own API key for Claude Opus 4

[originik]

Hello! I'm trying to set up Zed to use my own Anthrophic API key, but there just isn't enough information in the config documentation for me to know how to set everything up. Please help :)

Here is Zed's example configuration:

{
  "language_models": {
    "anthropic": {
      "available_models": [
        {
          "name": "claude-3-5-sonnet-20240620",
          "display_name": "Sonnet 2024-June",
          "max_tokens": 128000,
          "max_output_tokens": 2560,
          "cache_configuration": {
            "max_cache_anchors": 10,
            "min_total_token": 10000,
            "should_speculate": false
          },
          "tool_override": "some-model-that-supports-toolcalling"
        }
      ]
    }
  }
}

Here are the things I need help with:

1. The editor gives a warning for that configuration saying, "Missing property 'version'". I'm assuming that's the version of Anthropic that I'm using, but I don't know.
2. For max_tokens and max_output_tokens, should those be the maximum that the model supports or the maximum that I want Zed to use? If the latter, how should I make that decision?
3. I have no idea on what the three cache configuration options mean and I can't find any documentation on them.
4. tool_override: What is this and what should I put here? Assuming the same model would work here, should I consider putting a different model? I don't really know what this is used for, but if it is used for something that can be easily handled by a cheaper model, perhaps it would be better to use a cheaper model for the tool_override? I just really need help with this one.

Thanks in advance! I'm really excited to learn how to use AI to speed up development, but I'm very new to it all, so I'm looking for some guidance here so that I set things up correctly and don't waste my money.