Async by default
Submit a job, get a webhook, download results. No need to hold a connection.
24-hour SLA
Most jobs complete in 2-4 hours. Hard guarantee: 24 hours or the job is free.
50% of sync pricing
Same quality, same reliability, half the cost. The trade is time, not quality.
Up to 100K per job
Submit up to 100,000 requests per batch. Larger jobs queued automatically.
Standard format
JSONL input, JSONL output. Drop-in compatible with the OpenAI Batch API.
Persistent storage
Results stored for 30 days. Download via signed URL.
Pricing
Token-efficient, with volume discounts and burn incentives.
50% of sync rate
60% of sync rate
50% of sync rate
50% of sync rate
Pricing is illustrative. Final rates are governed by on-chain parameters and may vary based on network state.
Staking requirements
Tier-based access. Higher stakes unlock better economics and more capacity.
| Tier | Required stake | Access |
|---|---|---|
| Builder | 50M QUBIC | Batch API enabled |
| Startup | 150M QUBIC | Larger jobs, priority queue |
| Business | 500M QUBIC | Custom windows, dedicated capacity |
| Enterprise | Custom | On-prem, custom SLA |
Example
Drop-in compatible with the OpenAI SDK.
from aigarth import Aigarth
client = Aigarth(api_key="sk-aigarth-...")
# Submit a batch of classification requests
with open("requests.jsonl") as f:
job = client.batches.create(
input_file="file-abc123",
endpoint="/v1/chat/completions",
completion_window="24h",
)
# Wait for completion (or use webhook)
result = job.wait()
print(f"Processed {result.request_counts.completed} requests")Enterprise benefits
Everything in the standard tier, plus the things enterprises need.
- Same accuracy, half the cost
- Webhook callbacks for job completion
- 24-hour SLA or the job is free
- Results downloadable for 30 days
- Re-billable to internal cost centers
- Bulk processing at petabyte scale
- Migration from OpenAI Batch API in days
- Quarterly business reviews and volume discounts