LLMBOTTLENECK.COM
CohereLabs / cohere_compass_text

North-Micro-Vision-Instruct

2.48 billion parameters.

Open in the calculator →

Architecturepublished data
Quick answer
North-Micro-Vision-Instruct needs about 2.99 GB at Q4_K_M — the format most people download — with 8,192 tokens of context. It fits entirely on 18 of 18 common devices, starting with the NVIDIA GeForce RTX 5070 (12 GB). What else fits in 12 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 5070, 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.

Vision-language model. Sized for its text decoder; image inputs add memory not modelled here. Licence declared on the repository: apache-2.0.

Weight-file size by format

FormatSizeRangeBasis
FP164.98 GB3.89 GB – 5.00 GBreconstructed size
Q8_02.65 GB2.07 GB – 3.66 GBreconstructed size
Q6_K2.05 GB1.60 GB – 3.31 GBreconstructed size
Q5_K_M1.82 GB1.37 GB – 3.15 GBreconstructed size
Q5_01.79 GB1.34 GB – 3.13 GBreconstructed size
Q4_K_M1.60 GB1.15 GB – 3.00 GBreconstructed size
Q4_01.55 GB1.10 GB – 2.95 GBreconstructed size
Q3_K_M1.37 GB0.92 GB – 2.84 GBreconstructed size
Q2_K1.13 GB0.68 GB – 2.66 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 · 18 of 18 devices hold it entirely
DeviceMemoryVerdictDecodeCalculator
NVIDIA GeForce RTX 50702.99 GB of 12 GBfits~244 tok/sOpen →
AMD Radeon™ RX 9070 XT2.99 GB of 16 GBfits~266 tok/sOpen →
NVIDIA GeForce RTX 40802.99 GB of 16 GBfits~260 tok/sOpen →
NVIDIA GeForce RTX 50802.99 GB of 16 GBfits~348 tok/sOpen →
NVIDIA GeForce RTX 5070 Ti2.99 GB of 16 GBfits~325 tok/sOpen →
NVIDIA GeForce RTX 5060 Ti 16GB2.99 GB of 16 GBfits~162 tok/sOpen →
NVIDIA GeForce RTX 40902.99 GB of 24 GBfits~365 tok/sOpen →
AMD Radeon™ RX 7900 XTX2.99 GB of 24 GBfits~399 tok/sOpen →
NVIDIA GeForce RTX 30902.99 GB of 24 GBfits~339 tok/sOpen →
NVIDIA GeForce RTX 50902.99 GB of 32 GBfits~650 tok/sOpen →
NVIDIA RTX 6000 Ada Generation2.99 GB of 48 GBfits~348 tok/sOpen →
Apple M4 Pro2.99 GB of 64 GBfits~77 tok/sOpen →
Apple M5 Pro2.99 GB of 64 GBfits~82 tok/sOpen →
AMD Ryzen AI Max+ 395 with Radeon 8060S2.99 GB of 128 GBfits~106 tok/sOpen →
Apple M3 Max2.99 GB of 128 GBfits~94 tok/sOpen →
Apple M4 Max2.99 GB of 128 GBfits~107 tok/sOpen →
NVIDIA DGX Spark2.99 GB of 128 GBfits~99 tok/sOpen →
Apple M2 Ultra2.99 GB of 192 GBfits~122 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 North-Micro-Vision-Instruct

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
28 configurations in stock · 5 out of stock
MachineSpeedPer hourPer M tokensWhereRent
1× RTX 306012 GBrecommendedcheapestbest value~130 tok/sFaster than you read$0.036$0.076Vast.aiMarketplace · 99.8% reliableRent on Vast.ai
1× RTX 308010 GB~275 tok/sFaster than you read$0.082$0.083Vast.aiMarketplace · 98.3% reliableRent on Vast.ai
1× RTX 30708 GB~162 tok/sFaster than you read$0.082$0.14Vast.aiMarketplace · 99.1% reliableRunPod $0.13/h Rent on Vast.ai
1× RTX 4060 Ti 16GB16 GB~104 tok/sFaster than you read$0.090$0.24Vast.aiMarketplace · 99.4% reliableRent on Vast.ai
1× RTX 407012 GB~182 tok/sFaster than you read$0.096$0.15Vast.aiMarketplace · 99.8% reliableRent on Vast.ai
1× RTX 309024 GB~339 tok/sFaster than you read$0.12$0.10Vast.aiMarketplace · 99.5% reliableRunPod $0.50/h Rent on Vast.ai
1× RTX 5060 Ti 16GB16 GB~162 tok/sFaster than you read$0.12$0.21Vast.aiMarketplace · 99.1% reliableRent on Vast.ai
1× RTX 3080 Ti12 GB~330 tok/sFaster than you read$0.14$0.11Vast.aiMarketplace · 99.9% reliableRent 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 North-Micro-Vision-Instruct 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 North-Micro-Vision-Instruct 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 2026-09-13 at pinned commit 46b719694e3b; parameter count from the safetensors index at the same revision.

Architecture
cohere_compass_text
Layers
28
Hidden size
2,048
Attention heads
16
KV heads
8
Head dimension
128
Feed-forward width
6,144
Vocabulary
262,144
Context ceiling
500,000
Sliding window
4,096

Where the memory goes

tokenembedding× 28 decoder blocksGrouped-query attention16 query · 8 KV headswindow 4,096Feed-forwardone networkall activeoutputprojectiongrows with contextfixed per token

Grouped-query attention shares each key/value head across 2 query heads, so the KV cache is 50% 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 ↗

More from CohereLabs

North-Mini-Code-1.0

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":"coherelabs-north-micro-vision-instruct","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="coherelabs-north-micro-vision-instruct" 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. “North-Micro-Vision-Instruct VRAM and hardware requirements.” Architecture from CohereLabs/North-Micro-Vision-Instruct at revision 46b719694e3b, retrieved 2026-09-13. https://llmbottleneck.com/models/coherelabs-north-micro-vision-instruct

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