QuixiGate: Rust LLM gateway posts 8ms cold start and 24MiB RAM, ~29x LiteLLM JSON throughput
QuixiAI · x · 2026-11-14
QuixiAI released QuixiGate, an open-source LLM gateway written in Rust. It supports all major providers plus OpenRouter and OAuth subs for OpenAI and Grok, integrates OpenTelemetry and Arize Phoenix (you own your data), and ships as UI + headless. Benchmarks on an Apple M5 Max vs LiteLLM v1.109.1: cold start 8.4ms vs 4,224ms (503–1,972x faster), idle RSS 23.7MiB vs 941MiB (40–504x less RAM), stateless JSON at 70,690 req/s (29x LiteLLM's best 18-worker setup), and 974,742 tok/s across 1,000 governed GPU streams with zero errors.
More from coding & agent
- Claude Max hailed as 2026's best investment as Fable 'spawned thousands of startups' — EXM7777 · 2026-11-14
- ARC 1.0: A 0.6B Apache 2.0 Model That Routes AI Agents Across Cloud and Local Models — Balance- · 2026-11-14
- DevDay swag turned into Open.Dotlet, a 30-key physical prompting tablet for Codex — paw_lean · 2026-11-14
- NVIDIA open-sources SkillSpector: 26% of AI agent skills contain vulnerabilities — blaizedsouza · 2026-11-14
- Helium Solo: Free Tool to Record, Replay and Debug Failed AI Agents Locally — kasankilabs · 2026-11-14
- Dev Loves t3code So Much He Fears Anthropic Will Spend Millions to Beat It — adonis_singh · 2026-11-14