capy-flash

Two-GPU tensor parallel inference engine for Qwen3.8 models on consumer GPUs

Unclaimed

1 ratings4.34

Saves6

CategoryAI

PlatformWeb

PricingFree

capy-flash is an open-source two-GPU tensor-parallel large language model inference engine designed to run Qwen3.8-Flash-Next MoE and Qwen3.8-27B models efficiently on consumer-grade GPUs (2x16GB). It offloads experts to the CPU RAM for memory efficiency, supports NVFP4 quantization, and offers an OpenAI- and Anthropic-compatible server API for text generation and coding-agent workflows.

open-source

Ratings & Reviews

Write a review

4.34

1 reviews

Product

3.84

Design & Polish

4.84

Performance

4.34

Be the first to share a thoughtful review.

Information

CategoryAI

PlatformsWeb

PricingFree

AccountNot confirmed

ListedOctober 2026

WebVisit

Report a problem

Around the Web

Show HN6 source points · 0 comments