Kimi K3 goes live on Nebius with FP4 and claims 3–4x faster inference
teortaxesTex · x · 2026-07-27
Kimi K3 is now live on Nebius’s eu-north1 endpoint, with fp4 support.
The post highlights endpoint pricing and performance details: the same cost, but reportedly 3–4x faster. It’s effectively a deployment/update note for a frontier model hosted on a specific inference platform.
More from Infra
- TokenSpeed adds day-0 Kimi K3 support on NVIDIA Blackwell and AMD MI350X chips — zhyncs42 · 2026-07-27
- vLLM says K3’s 2.8T parameters were easy; caching them was the hard part — vllm_project · 2026-07-27
- Kimi K3 hardware estimates point to 480 GB VRAM and 8× H100 setups — cedric_chee · 2026-07-27
- Kimi K3 adds native MXFP4 quantization and shows heavy VRAM demands — teortaxesTex · 2026-07-27
- Baseten details day-0 API support for Moonshot's 2.8T-parameter Kimi K3 — philipkiely · 2026-07-27
- Builder launches a read-only MCP server for EU e-invoice preflight — jithox_AI · 2026-07-27