Last week, LiteLLM was compromised again. This encouraged me to post a little more about my project, because I believe that people are looking for an alternative right now.

GoModel is an open-source AI gateway OR a more efficient LiteLLM OR a self-hosted OpenRouter. It sits between your apps/agents/teams and LLM providers giving you one API for all of them + observability, budgets, guardrails, load balancing, failover, caching and more.

I started building it in December 2025 and improve it almost daily.

Numbers so far:

100K+ docker pulls

1K+ GitHub stars

6+ sponsors

25+ contributors

In comparison to LiteLLM:

25x lower p50 latency

consumes 20x less memory

16x smaller distroless docker image (~21MB) with a single binary is much easier to secure than a 350MB+ image full of Python dependencies

GitHub: https://github.com/ENTERPILOT/GoModel/

Benchmark (It's as fair as possible): https://github.com/ENTERPILOT/ai-gateway-reproducible-benchmark/

I'm sharing it mainly to make it more discoverable.…