Startup weighs ditching Claude Code: testing GLM via LiteLLM gateway to cut agent costs
Gonjanaenae319 · reddit · 2026-09-07
A startup built its managed-agent infrastructure around Claude Code but finds relying solely on Claude too costly. The poster has connected multiple models through a LiteLLM gateway smoothly and plans to try OpenCode, Pi, and DeepSeek Harness. They ask how much tool freedom enterprises have, whether teams mix frontier models with cheaper ones like Kimi or GLM, how these perform on large codebases, and how to evaluate moving off Claude Code to cut costs.
Related event: Startup ditches Claude Code over cost, tests GLM via LiteLLM gateway(2 posts)→
More from coding & agent
- MCP fault injection study: agent task scores drop 60% under realistic tool failures, scale doesn't help — dair_ai · 2026-10-26
- Codex manager pattern: one orchestrator agent directing three specialized peer agents — testingcatalog · 2026-10-26
- General agents burn ~140M tokens/month on LinkedIn outreach; scripts do it 60x cheaper — ChrisHarpon2 · 2026-10-26
- OpenAI upgrades Decisions API: 400 questions per call, native voice and image support — OpenAIDevs · 2026-10-26
- NLI alone rejects 5/6 new agent memories; a 0.45 similarity gate fixes it all — Alternative_Pin9598 · 2026-10-26
- LangChain ships Managed Deep Agents 0.9, walkthrough from Interrupt London keynote — LangChain · 2026-10-26