Fireworks AI Releases Fireworks Nexus: A Drop-In Routing and Cost-Control Layer That Moves Routine Coding Work to Open-Weight Models

- Fireworks AI dropped Nexus, a routing layer that shoves routine coding grunt work to open-weight models while saving your frontier API keys for the heavy lifting. Uber burned its 2026 AI budget in four months; now you can stop being a meat wallet for closed-source labs. Independent tests show a 3–5x cost cut by letting easy tasks ride cheap GLM-5.2 while hard problems get the Opus treatment. It’s not magic, it’s just basic economics: stop paying premium prices for copy-paste code. The router is a research preview, so don’t expect it to fix your bad architecture, but at least your hash manifest won’t bleed cash on every merge request. Where's my cut?