Batches

batch = client.create_batch({
    "model": "openai/gpt-5-mini",
    "prompts": [
        "Summarize this record.",
        "Classify this support ticket.",
    ],
    "system": "Be concise.",
    "max_tokens": 256,
    "completion_window": "24h",
    "session_id": "agent-run-42",
    "webhook_endpoint_id": "we_123",
})

status = client.get_batch("batch_123")
cancelling = client.cancel_batch("batch_123")
websocket_url = client.batches.websocket_url("batch_123", interval_ms=1500)

print(batch)
print(status)

​

Key parameters

• model: Model id used to infer the upstream provider. Most callers should not pass provider.
• prompts: Simple shorthand for one or more prompt strings.
• items: Structured prompt shorthand with id / custom_id, prompt, messages, per-row settings, or an advanced body.
• requests: Advanced inline provider-native batch rows.
• input_file_id: Existing file id uploaded through /batches/files?model=... for large prebuilt JSONL inputs.
• endpoint: Optional target endpoint for advanced rows. Phaseo chooses a model/provider default when omitted.
• completion_window: Upstream batch completion window such as 24h.
• session_id: Optional Phaseo grouping id for logs, sessions, and investigate tooling.
• webhook: Optional webhook configuration for async lifecycle notifications.
• metadata: Optional object persisted with the batch.