Kimi K3 with 1M context now usable in Claude Code via engy inference
richdotca · x · 2026-10-04
Inference provider engy now lets you run Kimi K3 with a 1M-token context window inside Claude Code by setting ANTHROPICBASEURL to engy's endpoint and ANTHROPICMODEL=kimi-k3[1m].
engy positions itself as a "verified inference" service for frontier open models with zero data retention and agent-native design. Its price list shows kimi-k3 at $3/$15 per million tokens (in/out), about 35% cheaper than the official API, alongside DeepSeek V4.1 Flash, GLM 5.3, and Qwen 3.8 endpoints.
More from coding & agent
- Bubblewrapping your AI coding session: sandboxing agents away from secrets and env files — Liu_eroteme · 2026-10-04
- Dev packages his creative-development opinions as an open-source Claude Code skill — jakubzeg · 2026-10-04
- Jev goes viral: five open-source agent projects from desktop control to drones and FPS games — blaizedsouza · 2026-10-04
- Claude Code weekly limits still hit within 3 days despite Opus 5.5, user reports 40k-token first message — Successful-Moment594 · 2026-10-04
- Matt Pocock ships skills v1.3 with /retro to mine old transcripts for improvements — mattpocockuk · 2026-10-04
- Free open-source LoRA trainer fits 9 models including MiniMax-H3 into 4GB consumer GPUs — AcademiaSD · 2026-10-04