PulseAugur
EN
LIVE 11:33:42

HyperQwen stack hits 118 tok/s on RTX 3090, but high-perf mode crashes

By PulseAugur Editorial · [1 sources] ·

A developer tested the HyperQwen stack on a single RTX 3090, achieving 118 tokens/second for chat and 237 tokens/second for verbatim context quoting. This performance significantly surpasses the 42 tokens/second baseline achieved with llama.cpp on the same model and hardware. However, the higher performance mode, which claims up to 381 tokens/second, crashed when attempting to process large context sizes, indicating instability in the speculative decoding implementation. AI

IMPACT Demonstrates achievable inference speeds on consumer GPUs, but highlights instability in advanced speculative decoding techniques.

RANK_REASON Developer testing and performance benchmarking of an LLM inference stack on consumer hardware.

Read on dev.to — LLM tag →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

HyperQwen stack hits 118 tok/s on RTX 3090, but high-perf mode crashes

How we ranked this

Signal score
38 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
Developer testing and performance benchmarking of an LLM inference stack on consumer hardware.
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
infra, product
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

Full methodology in our editorial standards.

COVERAGE [1]

  1. dev.to — LLM tag TIER_1 English(EN) · prodbymarcu ·

    I put the 381 tok/s Qwen stack on one RTX 3090, real chat does 118 and the headline mode crashes

    <p>A post went around earlier this month, 82 to 381 tok/s on a single RTX 3090 running a 27B model. I have that card and it's never idle, a coder model lives on it for my agent sessions, so I installed the stack and measured it on my own bench. Twice, because the first attempt di…

RELATED ENTITIES

RELATED TOPICS