Clarke 1.0

Sonnet-class inference. The production default for RAG, copilots and structured outputs

Clarke 1.0

Clarke 1.0 is the production default. It serves RAG pipelines, copilots, code assistance and structured outputs at the quality level of a mid-tier frontier model. Use Clarke for any workload where you need reliability, speed and cost in balance.

Model string

clarke-1.0

Specs

Spec	Value
Context window	1,000,000 tokens

Pricing is dynamic. For current rates see radium.cloud/pricing.

Capabilities

• Streaming: Supported
• Tool calling: Supported
• JSON mode: Supported
• JSON schema: Supported
• Reasoning: Supported
• Vision input: Supported

Recommended use cases

• RAG and retrieval-augmented generation
• Coding copilots and autocomplete
• Structured JSON output from natural language
• Tool-calling agents
• Production chat and customer support

Not suited for

Tasks that need deep reasoning over many steps, or very high-volume classification where every millisecond counts. Use Hal 1.0 or Tycho 1.0 for those.

Example request

cURL

curl -X POST https://api.radium.cloud/v1/chat/completions \\
  -H "Authorization: Bearer $RADIUM_API_KEY" \\
  -H "Content-Type: application/json" \\
  -d '{
    "model": "clarke-1.0",
    "messages": [{"role": "user", "content": "Summarize this API changelog."}],
    "response_format": {"type": "json_object"}
  }'