qwen38-27b-rtx3090
Serve large Qwen models fast on the GPUs you actually own. Qwen3.8-27B on a single 24 GB card with vLLM: 127 tok/s single-user (381 when the answer quotes the prompt), ~1,035 tok/s at 64 concurrent, 150k-262k context. vLLM patches, requant pipeline, benchmarks.
syv-ai/qwen38-27b-rtx3090
View on GitHub ↗Serve large Qwen models fast on the GPUs you actually own. Qwen3.8-27B on a single 24 GB card with vLLM: 127 tok/s single-user (381 when the answer quotes the prompt), ~1,035 tok/s at 64 concurrent, 150k-262k context. vLLM patches, requant pipeline, benchmarks.
qwen38-27b-rtx3090 grew +32.8% in the last 30 days, ranks 20th of 301 in Inference Optimization, actively maintained, last commit yesterday.
Star trajectory
64 days tracked+32.8% in 30d, above the Inference Optimization median of +1.5%.
Last commit yesterday, fresher than the median of 13d.
How it compares in Inference Optimization
Ranked 20th of 301 by stars.
- Language
- Python
- License
- Apache License 2.0
- Contributors
- 33
- Forks
- 295
- Watchers
- 19
- Open issues
- 3
- Releases
- 0
- Created
- 2026-08-15
- Stars this week
- +164
You've seen where qwen38-27b-rtx3090 stands today.
Get the moves before everyone else: the 5 Inference Optimization repos climbing fastest, every Tuesday.
Free weekly digest. Double opt-in, unsubscribe anytime.