Anthropic Claude models

Choose the right model in the family

Start with the workload and service target, then pick the smallest model that passes a representative evaluation set.

• Fable 5 targets the most demanding, long-running AI agent work.
• Opus 5 is the high-capability option for complex coding with AI agents and enterprise workflows.
• Sonnet 5 balances capability and speed for most production use; Haiku 4.5 is the faster, lower-cost option.

Inputs, outputs, context and tools

• Fable 5, Opus 5 and Sonnet 5 publish a 1M context window and 128K maximum output.
• Haiku 4.5 publishes a 200K context window and 64K maximum output.
• The current top-tier models use adaptive thinking; Haiku exposes extended thinking.
• Claude is especially relevant for coding, document analysis and multi-step tool workflows.

Where this family fits

These are strong starting points—not universal recommendations. Test them with your own prompts, tools and acceptance criteria.

• Software development and repository maintenance with AI agents.
• Policy, contract, and research-document analysis.
• Long-running tool workflows with explicit checkpoints.
• Classification and support flows that need Haiku's lower latency.

Reasoning, context and cost controls

• Track model lifecycle as a deployment dependency; retired IDs can fail without a code change.
• For Claude Code, use the console-generated Anthropic-style environment configuration.
• Separate official Anthropic credentials from gateway credentials and verify the destination before long runs.

Send a minimal request

This example uses the first provider model ID in the table. If Model Plaza shows a different ID for your API key, copy the Model Plaza value exactly.

curl https://app.llmfly.ai/v1/chat/completions \
  -H "Authorization: Bearer $LLMFLY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-fable-5",
    "messages": [{"role": "user", "content": "Explain this request in three bullets."}],
    "stream": false
  }'

Limits and compatibility notes

• Lifecycle dates are operational deadlines, not optional release notes.
• OpenAI-compatible examples do not automatically apply to Claude Code or Anthropic-native clients.
• Validate tool schemas, streaming events and thinking controls on the exact LLMFly AI route.

Production evaluation checklist

• Run 20–50 representative tasks and define pass criteria before testing.
• Measure time to first token, total latency, input and output tokens, tool retries, and cost per successful task.
• Test the endpoint, streaming mode, tool schemas, and structured output used by the application.
• Verify 400, 401, 404, 429, and temporary 5xx handling with bounded retries.
• Keep the model ID in configuration and validate a fallback before moving production requests.

What to confirm in Model Plaza

Model Plaza shows which model IDs your API key can use, their billing multipliers, availability, supported endpoints, and compatibility notes. Confirm these details with a small request before deployment.

• Claude Code uses Anthropic-style configuration rather than the OpenAI-compatible /v1 base URL.
• For Claude Code, copy the configuration shown by the console's Use Key action.
• Keep official and gateway credentials in separate terminal profiles to avoid sending traffic to the wrong account.