LLMBOTTLENECK.COM
Youssofal / qwen3_5_text

Qwen3.8-27B-MTPLX-Optimized-Speed

27B-class · 4-bit dynamic quantization.

The official card identifies this as a Qwen3.8-27B-based 4-bit dynamic-quantized MTPLX checkpoint. Its U32 storage total is a packed representation and the source metadata has a small internal dtype-total mismatch, so the UI uses the official 27B class label rather than a falsely precise logical total. Open the official parameter note ↗

Open in the calculator →

Architecturepublished data
Quick answer
Qwen3.8-27B-MTPLX-Optimized-Speed needs about 18.1 GB at Q4_K_M — the format most people download — with 8,192 tokens of context. It fits entirely on 12 of 18 common devices, starting with the NVIDIA GeForce RTX 4090 (24 GB). What else fits in 24 GB. No hardware for it? Rent a GPU that runs it, priced live.

Sized at 8,192 tokens of context with the whole model in device memory, across 18 common devices. Speeds are estimates, not benchmarks.

Find the NVIDIA GeForce RTX 4090, the smallest card listed here that holds it: Amazon ↗ · eBay (new and used) ↗Store links may pay us a commission. They never decide which card is suggested — the memory arithmetic does.

Weight-file size by format

FormatSizeRangeBasis
FP1654.0 GB37.8 GB – 54.5 GBsize range
Q8_028.7 GB20.1 GB – 31.4 GBsize range
Q6_K22.2 GB15.5 GB – 25.4 GBsize range
Q5_K_M19.3 GB13.0 GB – 25.4 GBsize range
Q5_018.8 GB13.0 GB – 25.4 GBsize range
Q4_K_M16.6 GB10.6 GB – 25.4 GBsize range
Q4_015.7 GB10.6 GB – 25.4 GBsize range
Q3_K_M13.5 GB8.12 GB – 25.4 GBsize range
Q2_K10.7 GB6.20 GB – 25.4 GBsize range

A file size is not the memory a run needs: the KV cache and the runtime reserve come on top, and the calculator adds both.

What the labels mean
published data
Read from a published source — a file's byte count, a model's configuration or a manufacturer's specification — or exact arithmetic on such values. Not a measurement on a machine.
reconstructed size
Weight size reconstructed from the pinned architecture, because no published file exists.
size range
Only a lower and an upper bound are claimed for this weight size.

Hardware ladder

Q4_K_M at 8,192 tokens · 12 of 18 devices hold it entirely
DeviceMemoryVerdictDecodeCalculator
NVIDIA GeForce RTX 507018.1 GB of 12 GB24 layers on system RAM~9.2 tok/s with offloadOpen →
AMD Radeon™ RX 9070 XT18.1 GB of 16 GB8 layers on system RAM~18 tok/s with offloadOpen →
NVIDIA GeForce RTX 408018.1 GB of 16 GB8 layers on system RAM~18 tok/s with offloadOpen →
NVIDIA GeForce RTX 508018.1 GB of 16 GB8 layers on system RAM~20 tok/s with offloadOpen →
NVIDIA GeForce RTX 5070 Ti18.1 GB of 16 GB8 layers on system RAM~20 tok/s with offloadOpen →
NVIDIA GeForce RTX 5060 Ti 16GB18.1 GB of 16 GB8 layers on system RAM~14 tok/s with offloadOpen →
NVIDIA GeForce RTX 409018.1 GB of 24 GBfits~46 tok/sOpen →
AMD Radeon™ RX 7900 XTX18.1 GB of 24 GBfits~51 tok/sOpen →
NVIDIA GeForce RTX 309018.1 GB of 24 GBfits~43 tok/sOpen →
NVIDIA GeForce RTX 509018.1 GB of 32 GBfits~82 tok/sOpen →
NVIDIA RTX 6000 Ada Generation18.1 GB of 48 GBfits~44 tok/sOpen →
Apple M4 Pro18.1 GB of 64 GBfits~16 tok/sAbout reading paceOpen →
Apple M5 Pro18.1 GB of 64 GBfits~17 tok/sAbout reading paceOpen →
AMD Ryzen AI Max+ 395 with Radeon 8060S18.1 GB of 128 GBfits~13 tok/sAbout reading paceOpen →
Apple M3 Max18.1 GB of 128 GBfits~22 tok/sAbout reading paceOpen →
Apple M4 Max18.1 GB of 128 GBfits~29 tok/sAbout reading paceOpen →
NVIDIA DGX Spark18.1 GB of 128 GBfits~13 tok/sAbout reading paceOpen →
Apple M2 Ultra18.1 GB of 192 GBfits~39 tok/sOpen →

Decode is a calibrated estimate from the published calibration; capacity is the manufacturer's published ceiling, not guaranteed free memory. Fitting in memory is not the same as loading: whether the runtime and version you have supports this architecture and format on that machine has not been tested here. Offloaded rows assume there is enough system RAM for the overflow — the calculator checks that against the RAM you declare. Each “Open” link keeps this model, format and context.

Run it in the cloud

Rent a GPU that runs Qwen3.8-27B-MTPLX-Optimized-Speed

Every rentable machine that holds the whole model, from Vast.ai and RunPod, with this site's speed estimate for each and the provider's own price, read live. Cheapest first; sort by cost per token or speed instead, or switch the format.

Prices · 18:30 UTC, 22 Sept
27 configurations in stock · 6 out of stock
MachineSpeedPer hourPer M tokensWhereRent
2× RTX 306024 GBcheapest~32 tok/sFaster than you read$0.11$0.93Vast.aiMarketplace · 98.0% reliableRent on Vast.ai
1× RTX 309024 GBrecommendedbest value~43 tok/sFaster than you read$0.12$0.79Vast.aiMarketplace · 99.5% reliableRunPod $0.50/h Rent on Vast.ai
1× RTX 3090 Ti24 GB~46 tok/sFaster than you read$0.19$1.12Vast.aiMarketplace · 99.6% reliableRunPod $0.27/h Rent on Vast.ai
2× RTX 407024 GB~45 tok/sFaster than you read$0.19$1.15Vast.aiMarketplace · 99.8% reliableRent on Vast.ai
2× RTX 4060 Ti 16GB32 GB~26 tok/sAbout reading pace$0.22$2.38Vast.aiMarketplace · 99.7% reliableRent on Vast.ai
1× L424 GB~13 tok/sAbout reading pace$0.26$5.14Vast.aiMarketplace · 98.9% reliableRunPod $0.49/h Rent on Vast.ai
2× RTX 5060 Ti 16GB32 GB~40 tok/sFaster than you read$0.26$1.74Vast.aiMarketplace · 99.5% reliableRent on Vast.ai
1× RTX 6000 Ada48 GB~44 tok/sFaster than you read$0.26$1.64Vast.aiMarketplace · 99.9% reliableRunPod $0.84/h Rent on Vast.ai

First time renting a GPU? How it works, in four steps
  1. Create an account and add credit. Both providers are prepaid: $10 is enough to try any card on this page for hours. A new RunPod account that arrives through a referral link gets a one-time $5 credit when it first adds $10.
  2. Pick the card and the count shown here, and a template: llama.cpp or vLLM for an OpenAI-compatible API, or one with a web chat if you only want to talk to the model. “How to launch it” below gives the exact image and command.
  3. Wait for the download. The weights are fetched on the machine; tens of gigabytes take a few minutes on a data-centre connection. You pay by the second from the moment it starts.
  4. Stop it when you are done. A running machine bills even when idle. A stopped one costs nothing per hour, though its disk is usually still billed until you delete it.
How to launch it

No Q4_K_M file of Qwen3.8-27B-MTPLX-Optimized-Speed is catalogued here, so there is no exact command to vouch for. Start the machine from the llama.cpp server image ghcr.io/ggml-org/llama.cpp:server-cuda and point -hf at a Q4_K_M build from the publisher or a quantizer you trust — search Hugging Face for one. Check its file size against the memory figure on this page before you rent.

Every machine holds the whole model at Q4_K_M and 8,192 tokens of context, no offload. Speeds are this site’s single-stream decode estimates; prices are what each provider’s own API quoted, on-demand, for the whole machine. Vast hosts below 98% measured reliability are left out.

Referral links Vast.ai, RunPod and Novita pay us a share of what you spend if you sign up through these buttons. It costs you nothing, and it never decides an order or a recommendation: both are computed from the live price and the speed, and options that pay us nothing are listed and recommended on the same terms. How we rank

Buying a card instead, or paying by the token? Compare a year of Qwen3.8-27B-MTPLX-Optimized-Speed three ways →

Under the hood

Architecture, read from the publisher’s file

The numbers every figure above is computed from, with the file they came from.

Architecture

✓ Architecture read from the published config.json

Retrieved 2026-09-01 at pinned commit 123db8bcc710; the native packed/fused checkpoint metadata is not used as a logical parameter total.

Architecture
qwen3_5_text
Layers
64
Hidden size
5,120
Attention heads
24
KV heads
4
Head dimension
256
Feed-forward width
17,408
Vocabulary
248,320
Context ceiling
262,144
RoPE theta
10,000,000

Where the memory goes

tokenembedding× 64 decoder blocksGrouped-query attention24 query · 4 KV headsfull contextFeed-forwardone networkall activeoutputprojectiongrows with contextfixed per token

Grouped-query attention shares each key/value head across 6 query heads, so the KV cache is 17% of what multi-head attention would need at the same context.

This is a multimodal checkpoint (vision); the diagram and memory sizing above cover the text decoder. Encoder/audio/image activations are not included in the KV or weight figures.

Open the exact official config.json ↗

Use this answer in your own product

As JSON, from the API
curl -s https://llmbottleneck.com/v1/analyze \
  -H "Authorization: Bearer $LLMB_KEY" -H "content-type: application/json" \
  -d '{"model":"youssofal-qwen3-8-27b-mtplx-optimized-speed","quantization":"Q4_K_M","context":8192,"hardware":"nvidia-geforce-rtx-4090"}'

Same engine, same evidence, every field sourced. Free key in one step →

On your page, as a widget
<script src="https://llmbottleneck.com/widget.js"
  data-model="youssofal-qwen3-8-27b-mtplx-optimized-speed" data-quantization="Q4_K_M"
  data-hardware="nvidia-geforce-rtx-4090" data-context="8192"></script>

No key needed. Unbranded, with your own buy button, on Pro and Business →

Citing this page

LLM Bottleneck. “Qwen3.8-27B-MTPLX-Optimized-Speed VRAM and hardware requirements.” Architecture from Youssofal/Qwen3.8-27B-MTPLX-Optimized-Speed at revision 123db8bcc710, retrieved 2026-09-01. https://llmbottleneck.com/models/youssofal-qwen3-8-27b-mtplx-optimized-speed

Every figure above is either the published value or a reconstruction whose measured error is on the accuracy page.

llmbottleneck
catalogue 2026-10-03models 327devices 135