LLMBOTTLENECK.COM
inclusionAI / bailing_hybrid

Ling-3.0-flash

127.5 billion parameters, routing 8 of 512 experts per token.

Open in the calculator →

Architecturepublished data
Quick answer
Ling-3.0-flash needs about 78.2 GB at Q4_K_M — the format most people download — with 8,192 tokens of context. It fits entirely on 5 of 18 common devices, starting with the AMD Ryzen AI Max+ 395 with Radeon 8060S (128 GB). What else fits in 128 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 AMD Ryzen AI Max+ 395 with Radeon 8060S, 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
FP16255.1 GBexactpublished data
Q8_0135.6 GBexactpublished data
Q6_K104.8 GBexactpublished data
Q5_K_M90.5 GBexactpublished data
Q5_089.0 GB61.4 GB – 106.6 GBsize range
Q4_K_M77.0 GBexactpublished data
Q4_074.2 GB50.2 GB – 106.6 GBsize range
Q3_K_M63.8 GB38.3 GB – 106.6 GBsize range
Q2_K50.5 GB29.3 GB – 106.6 GBsize range

5 of these are published files. Open the pinned GGUF repository ↗

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 · 5 of 18 devices hold it entirely
DeviceMemoryVerdictDecodeCalculator
NVIDIA GeForce RTX 507078.2 GB of 12 GB37 layers on system RAM~25 tok/s with offloadOpen →
AMD Radeon™ RX 9070 XT78.2 GB of 16 GB34 layers on system RAM~27 tok/s with offloadOpen →
NVIDIA GeForce RTX 408078.2 GB of 16 GB34 layers on system RAM~27 tok/s with offloadOpen →
NVIDIA GeForce RTX 508078.2 GB of 16 GB34 layers on system RAM~27 tok/s with offloadOpen →
NVIDIA GeForce RTX 5070 Ti78.2 GB of 16 GB34 layers on system RAM~27 tok/s with offloadOpen →
NVIDIA GeForce RTX 5060 Ti 16GB78.2 GB of 16 GB34 layers on system RAM~26 tok/s with offloadOpen →
NVIDIA GeForce RTX 409078.2 GB of 24 GB30 layers on system RAM~30 tok/s with offloadOpen →
AMD Radeon™ RX 7900 XTX78.2 GB of 24 GB30 layers on system RAM~30 tok/s with offloadOpen →
NVIDIA GeForce RTX 309078.2 GB of 24 GB30 layers on system RAM~30 tok/s with offloadOpen →
NVIDIA GeForce RTX 509078.2 GB of 32 GB26 layers on system RAM~35 tok/s with offloadOpen →
NVIDIA RTX 6000 Ada Generation78.2 GB of 48 GB17 layers on system RAM~48 tok/s with offloadOpen →
Apple M4 Pro78.2 GB of 64 GBdoes not fitnot calibratedOpen →
Apple M5 Pro78.2 GB of 64 GBdoes not fitnot calibratedOpen →
AMD Ryzen AI Max+ 395 with Radeon 8060S78.2 GB of 128 GBfits~65 tok/sOpen →
Apple M3 Max78.2 GB of 128 GBfits~72 tok/sOpen →
Apple M4 Max78.2 GB of 128 GBfits~86 tok/sOpen →
NVIDIA DGX Spark78.2 GB of 128 GBfits~61 tok/sOpen →
Apple M2 Ultra78.2 GB of 192 GBfits~102 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 Ling-3.0-flash

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
18 configurations in stock · 13 out of stock
MachineSpeedPer hourPer M tokensWhereRent
8× RTX 306096 GBcheapest~335 tok/sFaster than you read$0.59$0.49Vast.aiMarketplace · 99.7% reliableRent on Vast.ai
4× RTX 309096 GB~599 tok/sFaster than you read$0.60$0.28Vast.aiMarketplace · 99.4% reliableRent on Vast.ai
4× RTX 5090128 GBrecommendedbest value~1147 tok/sFaster than you read$1.07$0.26Vast.aiMarketplace · 99.6% reliableRent on Vast.ai
2× RTX 6000 Ada96 GB~377 tok/sFaster than you read$1.07$0.79Vast.aiMarketplace · 99.4% reliableRent on Vast.ai
1× RTX PRO 600096 GB~398 tok/sFaster than you read$1.14$0.79Vast.aiMarketplace · 98.8% reliableRunPod $1.69/h Rent on Vast.ai
1× A100 80GB PCIe80 GB~430 tok/sFaster than you read$1.19$0.77RunPodCommunity CloudRent on RunPod
2× RTX A600096 GB~302 tok/sFaster than you read$1.20$1.11Vast.aiMarketplace · 99.4% reliableRent on Vast.ai
4× L496 GB~192 tok/sFaster than you read$1.29$1.86Vast.aiMarketplace · 99.4% reliableRent on Vast.ai

No GPU at all

Use Ling-3.0-flash by the token

For one person chatting, that is about 4.1× cheaper than the best-value rented card above ($0.26 per million tokens). A rented GPU pays off when you keep it busy — many requests batched together, long agent runs — or when the data must stay on a machine you control, or the exact file you want is not served anywhere.

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 Q4_K_M on it (llama.cpp)
Container image
ghcr.io/ggml-org/llama.cpp:server-cuda
Start command / arguments
-hf inclusionAI/Ling-3.0-flash-GGUF:Q4_K_M -c 8192 -ngl 999 --host 0.0.0.0 --port 8080
Expose port
8080 — an OpenAI-compatible API at /v1

Starting from a saved llama.cpp template instead? Leave the arguments empty and set these variables — only the first two change between models:

LLAMA_ARG_HF_REPO=inclusionAI/Ling-3.0-flash-GGUF:Q4_K_M
LLAMA_ARG_CTX_SIZE=8192
LLAMA_ARG_N_GPU_LAYERS=999
LLAMA_ARG_HOST=0.0.0.0
LLAMA_ARG_PORT=8080

Already have the runtime? llama-server -hf inclusionAI/Ling-3.0-flash-GGUF:Q4_K_M -c 8192 -ngl 999 --host 0.0.0.0 --port 8080. Weights: inclusionAI/Ling-3.0-flash-GGUF. The CUDA image is for NVIDIA cards; an AMD machine needs llama.cpp's ROCm build.

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 Ling-3.0-flash 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 42766a814ab1; parameter count from the safetensors index at the same revision.

Architecture
bailing_hybrid
Layers
42
Hidden size
2,560
Attention heads
32
KV heads
32
Head dimension
128
Feed-forward width
6,144
Vocabulary
157,184
Context ceiling
262,144
RoPE theta
6,000,000
Experts
512
Experts / token
8
Expert width
768
Shared experts
1
Shared expert width
768
Latent KV rank
512

Where the memory goes

tokenembedding× 42 decoder blocksLatent attentioncompressed KV rankfull contextRouted experts8 of 512 per tokenall residentoutputprojectiongrows with contextcapacity ≠ traffic

Latent attention projects keys and values into a compressed rank before caching them, which is why its KV cache is a fraction of a comparable dense model.

Open the exact official config.json ↗

More from inclusionAI

Ling-3.0-tinyLLaDA2.0-miniRing-2.5-1T

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":"inclusionai-ling-3-0-flash","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="inclusionai-ling-3-0-flash" 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. “Ling-3.0-flash VRAM and hardware requirements.” Architecture from inclusionAI/Ling-3.0-flash at revision 42766a814ab1, retrieved 2026-09-01. https://llmbottleneck.com/models/inclusionai-ling-3-0-flash

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