@TeksEdge: A localmaxxer hit ~381 tok/s on a SINGLE RTX 3090 with Qwen3.8-27B. This developer has turned a 24GB RTX 3090 into a mo…

X AI KOLs Timeline News

Summary

A developer achieved up to 381 tok/s inference speed on a single RTX 3090 with the Qwen3.8-27B model using optimized techniques like DFlash2 and prefix caching, particularly effective for document-based tasks like RAG and coding assistants.

A localmaxxer hit ~381 tok/s on a SINGLE RTX 3090 with Qwen3.8-27B. This developer has turned a 24GB RTX 3090 into a monster Qwen inference box - w/some creativity. Four days ago ~82 tok/s single-user Then ~114 tok/s with optimized MTP ~138 tok/s with DFlash2 + lookup drafting Now ~381 tok/s on ONE request How? The recipe combines ... Qwen3.8-27B 1× RTX 3090 24GB @ 250W heavily optimized vLLM DFlash2 speculative decoding lookup-augmented drafting prefix caching 16-token verification blocks quantized KV / heads / activations DFlash2 normally proposes 7 tokens. The developer realized the verification block doesn't have to stop there. If Qwen is answering from a document already sitting in the prompt, the system can fill the remaining draft positions using tokens found directly in that context. So the target model can verify 16 tokens at once. On a ~25K-token document reproduction task: Previous DFlash2 ~260 tok/s Longer verification + context lookup ~382 tok/s Acceptance: 15 of 16 tokens per verification step That is where the crazy number comes from. On ordinary real-world chat prompts, the same setup is around ~133 tok/s Still extremely fast for a dense 27B model on an RTX 3090. The 381 tok/s mode shines when the answer largely comes from material already in context so these are best use cases RAG / document Q&A Coding assistants applying edits Quoting or rewriting documents Extracting information from long prompts And another optimization With prefix caching, a second question against the same 25K-token document reportedly goes from: 22.4 sec TTFT → 0.56 sec TTFT Because the model doesn't need to process the whole document again. It's specifically a mode for RAG front ends and coding agents. Follow iamMess on Reddit or syv-ai on GitHub Reddit: r/LocalLLaMA/comments/1vtup5s/ GitHub: /syv-ai/qwen38-27b-rtx3090
Original Article
View Cached Full Text

Cached at: 08/21/26, 03:19 PM

A localmaxxer hit ~381 tok/s on a SINGLE RTX 3090 with Qwen3.8-27B.

This developer has turned a 24GB RTX 3090 into a monster Qwen inference box - w/some creativity.

Four days ago ~82 tok/s single-user

Then ~114 tok/s with optimized MTP ~138 tok/s with DFlash2 + lookup drafting

Now ~381 tok/s on ONE request

How? The recipe combines … Qwen3.8-27B 1× RTX 3090 24GB @ 250W heavily optimized vLLM DFlash2 speculative decoding lookup-augmented drafting prefix caching 16-token verification blocks quantized KV / heads / activations

DFlash2 normally proposes 7 tokens.

The developer realized the verification block doesn’t have to stop there.

If Qwen is answering from a document already sitting in the prompt, the system can fill the remaining draft positions using tokens found directly in that context.

So the target model can verify 16 tokens at once.

On a ~25K-token document reproduction task: Previous DFlash2 ~260 tok/s

Longer verification + context lookup ~382 tok/s

Acceptance: 15 of 16 tokens per verification step

That is where the crazy number comes from.

On ordinary real-world chat prompts, the same setup is around ~133 tok/s

Still extremely fast for a dense 27B model on an RTX 3090.

The 381 tok/s mode shines when the answer largely comes from material already in context so these are best use cases RAG / document Q&A Coding assistants applying edits Quoting or rewriting documents Extracting information from long prompts

And another optimization With prefix caching, a second question against the same 25K-token document reportedly goes from: 22.4 sec TTFT → 0.56 sec TTFT

Because the model doesn’t need to process the whole document again.

It’s specifically a mode for RAG front ends and coding agents.

Follow iamMess on Reddit or syv-ai on GitHub

Reddit: r/LocalLLaMA/comments/1vtup5s/ GitHub: /syv-ai/qwen38-27b-rtx3090

Similar Articles