LLMBOTTLENECK.COM
Qwen / qwen3_5_moe_text

Qwen3.8-2.4T-A95B

2446.2 billion parameters, routing 10 of 512 experts per token.

Open in the calculator →

Architecturepublished data
Quick answer
Qwen3.8-2.4T-A95B needs about 1459.7 GB at Q4_K_M — the format most people download — with 8,192 tokens of context. None of the 18 common devices listed below holds it entirely at this setting. 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.

Weight-file size by format

FormatSizeRangeBasis
FP164813.1 GB4710.2 GB – 4915.9 GBreconstructed size
Q8_02557.7 GB2503.0 GB – 2612.3 GBreconstructed size
Q6_K1975.0 GB1932.8 GB – 2017.2 GBreconstructed size
Q5_K_M1708.4 GB1671.9 GB – 1744.9 GBreconstructed size
Q5_01655.8 GB1620.4 GB – 1691.2 GBreconstructed size
Q4_K_M1457.5 GB1426.4 GB – 1488.6 GBreconstructed size
Q4_01355.3 GB1326.4 GB – 1384.3 GBreconstructed size
Q3_K_M1190.9 GB1088.0 GB – 1293.7 GBreconstructed size
Q2_K872.4 GB823.3 GB – 921.6 GBreconstructed size

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 · 0 of 18 devices hold it entirely
DeviceMemoryVerdictDecodeCalculator
NVIDIA GeForce RTX 50701459.7 GB of 12 GB92 layers on system RAM~1.2 tok/s with offloadOpen →
AMD Radeon™ RX 9070 XT1459.7 GB of 16 GB92 layers on system RAM~1.2 tok/s with offloadOpen →
NVIDIA GeForce RTX 40801459.7 GB of 16 GB92 layers on system RAM~1.2 tok/s with offloadOpen →
NVIDIA GeForce RTX 50801459.7 GB of 16 GB92 layers on system RAM~1.2 tok/s with offloadOpen →
NVIDIA GeForce RTX 5070 Ti1459.7 GB of 16 GB92 layers on system RAM~1.2 tok/s with offloadOpen →
NVIDIA GeForce RTX 5060 Ti 16GB1459.7 GB of 16 GB92 layers on system RAM~1.2 tok/s with offloadOpen →
NVIDIA GeForce RTX 40901459.7 GB of 24 GB91 layers on system RAM~1.2 tok/s with offloadOpen →
AMD Radeon™ RX 7900 XTX1459.7 GB of 24 GB91 layers on system RAM~1.2 tok/s with offloadOpen →
NVIDIA GeForce RTX 30901459.7 GB of 24 GB91 layers on system RAM~1.2 tok/s with offloadOpen →
NVIDIA GeForce RTX 50901459.7 GB of 32 GB91 layers on system RAM~1.2 tok/s with offloadOpen →
NVIDIA RTX 6000 Ada Generation1459.7 GB of 48 GB90 layers on system RAM~1.2 tok/s with offloadOpen →
Apple M4 Pro1459.7 GB of 64 GBdoes not fitnot calibratedOpen →
Apple M5 Pro1459.7 GB of 64 GBdoes not fitnot calibratedOpen →
AMD Ryzen AI Max+ 395 with Radeon 8060S1459.7 GB of 128 GBdoes not fitnot calibratedOpen →
Apple M3 Max1459.7 GB of 128 GBdoes not fitnot calibratedOpen →
Apple M4 Max1459.7 GB of 128 GBdoes not fitnot calibratedOpen →
NVIDIA DGX Spark1459.7 GB of 128 GBdoes not fitnot calibratedOpen →
Apple M2 Ultra1459.7 GB of 192 GBdoes not fitnot calibratedOpen →

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 Qwen3.8-2.4T-A95B

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

The machines that hold Qwen3.8-2.4T-A95B at Q4_K_M are not in stock at Vast.ai or RunPod right now.

No GPU at all

Use Qwen3.8-2.4T-A95B by the token

Nothing to set up and nothing to switch off: you pay only for the tokens you use.

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 Qwen3.8-2.4T-A95B 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 207bd685a7e3; parameter count from the safetensors index at the same revision.

Architecture
qwen3_5_moe_text
Layers
92
Hidden size
8,192
Attention heads
64
KV heads
4
Head dimension
256
Vocabulary
248,320
Context ceiling
262,144
RoPE theta
10,000,000
Experts
512
Experts / token
10
Expert width
2,048
Shared expert width
2,048

Where the memory goes

tokenembedding× 92 decoder blocksGrouped-query attention64 query · 4 KV headsfull contextRouted experts10 of 512 per tokenall residentoutputprojectiongrows with contextcapacity ≠ traffic

Grouped-query attention shares each key/value head across 16 query heads, so the KV cache is 6% of what multi-head attention would need at the same context.

Open the exact official config.json ↗

More from Qwen

Qwen-AgentWorld-35B-A3BQwen2.5-0.5BQwen2.5-0.5B-InstructQwen2.5-1.5B-InstructQwen2.5-14B-InstructQwen2.5-32B-InstructQwen2.5-3B-InstructQwen2.5-72B-Instruct

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":"qwen-qwen3-8-2-4t-a95b","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="qwen-qwen3-8-2-4t-a95b" 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. “Qwen3.8-2.4T-A95B VRAM and hardware requirements.” Architecture from Qwen/Qwen3.8-2.4T-A95B at revision 207bd685a7e3, retrieved 2026-09-01. https://llmbottleneck.com/models/qwen-qwen3-8-2-4t-a95b

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