Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
The Hybrid Edge: How SLMs and Intent Routing Cut My AI Spend by 70%
Quick Answer (2026): I cut inference spend by ~70% by routing 80–90% of requests to a local SLM and using a frontier API only for hard cases. Pricing is bipolar now: “commodity cheap” (DeepSeek V3.2) vs “frontier premium” (GPT-5.2, Claude…