Paste your xAI API key. Grok-4 calls flow through aiusage's cache + cost-optimizer. Same runs balance that pays for Claude + GPT. Every model, one dashboard.
Paste your xAI key on the dashboard. Encrypted on arrival (AES-256-GCM, zero-knowledge).
Point your client at aiusage.ai/xai/v1. Done.
Grok-4, Grok-4-fast, Grok-code โ all cached + cost-optimized. Mix with Claude + GPT freely.
Billed from your aiusage runs balance. xAI's key pays upstream; we take the cache + routing margin and hand you the savings.
Grok is fastest for real-time / search-heavy tasks. Claude wins at long-context code. GPT dominates reasoning benchmarks. Teams running all three end up with three dashboards, three caches, three invoices. aiusage collapses all of it into one meter.
| Usage | Direct xAI | Through aiusage | You save |
|---|---|---|---|
| 1K Grok-4 calls/mo | ~$30 | ~$12 | ~$18 |
| 10K Grok-4 calls/mo | ~$150 | ~$60 | ~$90 |
| All 3 providers blended | ~$850 | ~$340 | ~$510 |
No subscription. Mix providers in one call path โ aiusage routes to the cheapest viable model for each prompt.