KVarN
KVarN is a native vLLM KV-cache quantization backend for your agents: 3-5x more context, throughput above FP16, and FP16-level accuracy. Calibratio...
Updated means this listing was last refreshed on Jun 16, 2026.
About KVarN
Pricing:Freemium
KVarN is a native vLLM KV-cache quantization backend for your agents: 3-5x more context, throughput above FP16, and FP16-level accuracy. Calibration-free, one flag. huawei-csl/KVarN is an open source project on GitHub with 399 stars. Built primarily in Python. Licensed under Apache-2.0. Topics: agentic-ai, kv-cache, llm, llm-inference, long-context, quantization, vllm.
Alternatives
Claude Opus 4.7
87.6% SWE-bench — the strongest coding model available
