NousResearch / llama

Llama-2-7b-hf

Name: Llama-2-7b-hf architecture and memory requirements
Creator: LLM Bottleneck
License: https://creativecommons.org/licenses/by/4.0/

6.74 billion parameters.

Open in the calculator →

Architecturepublished data

Quick answer

Llama-2-7b-hf needs about 7.03 GB at Q4_K_M — the format most people download — with 4,096 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.

Smallest device that fitsNVIDIA GeForce RTX 507012 GB · 672 GB/s~74 tok/s Faster than you readOpen in the calculator →
Fastest in this listNVIDIA GeForce RTX 509032 GB · 1792 GB/s~198 tok/s Faster than you readOpen in the calculator →

Sized at 4,096 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.

Weight-file size by format

every format, with its evidence

Format	Size	Range	Basis
FP16	13.5 GB	13.2 GB – 13.5 GB	reconstructed size
Q8_0	7.16 GB	7.02 GB – 7.20 GB	reconstructed size
Q6_K	5.53 GB	5.42 GB – 5.56 GB	reconstructed size
Q5_K_M	4.78 GB	4.67 GB – 4.81 GB	reconstructed size
Q5_0	4.65 GB	4.54 GB – 4.68 GB	reconstructed size
Q4_K_M	4.08 GB	3.97 GB – 4.12 GB	reconstructed size
Q4_0	3.83 GB	3.72 GB – 3.86 GB	reconstructed size
Q3_K_M	3.34 GB	3.10 GB – 3.57 GB	reconstructed size
Q2_K	2.48 GB	2.37 GB – 2.58 GB	reconstructed 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 4,096 tokens · 18 of 18 devices hold it entirely

Device	Memory	Verdict	Decode	Calculator
NVIDIA GeForce RTX 5070	7.03 GB of 12 GB	fits	~74 tok/s	Open →
AMD Radeon™ RX 9070 XT	7.03 GB of 16 GB	fits	~81 tok/s	Open →
NVIDIA GeForce RTX 4080	7.03 GB of 16 GB	fits	~79 tok/s	Open →
NVIDIA GeForce RTX 5080	7.03 GB of 16 GB	fits	~106 tok/s	Open →
NVIDIA GeForce RTX 5070 Ti	7.03 GB of 16 GB	fits	~99 tok/s	Open →
NVIDIA GeForce RTX 5060 Ti 16GB	7.03 GB of 16 GB	fits	~49 tok/s	Open →
NVIDIA GeForce RTX 4090	7.03 GB of 24 GB	fits	~111 tok/s	Open →
AMD Radeon™ RX 7900 XTX	7.03 GB of 24 GB	fits	~121 tok/s	Open →
NVIDIA GeForce RTX 3090	7.03 GB of 24 GB	fits	~103 tok/s	Open →
NVIDIA GeForce RTX 5090	7.03 GB of 32 GB	fits	~198 tok/s	Open →
NVIDIA RTX 6000 Ada Generation	7.03 GB of 48 GB	fits	~106 tok/s	Open →
Apple M4 Pro	7.03 GB of 64 GB	fits	~34 tok/s	Open →
Apple M5 Pro	7.03 GB of 64 GB	fits	~37 tok/s	Open →
AMD Ryzen AI Max+ 395 with Radeon 8060S	7.03 GB of 128 GB	fits	~32 tok/s	Open →
Apple M3 Max	7.03 GB of 128 GB	fits	~45 tok/s	Open →
Apple M4 Max	7.03 GB of 128 GB	fits	~56 tok/s	Open →
NVIDIA DGX Spark	7.03 GB of 128 GB	fits	~30 tok/s	Open →
Apple M2 Ultra	7.03 GB of 192 GB	fits	~72 tok/s	Open →

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 Llama-2-7b-hf

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

Machine	Speed	Per hour	Per M tokens	Where	Rent
1× RTX 306012 GBrecommendedcheapestbest value	~39 tok/sFaster than you read	$0.036	$0.25	Vast.aiMarketplace · 99.8% reliable	Rent on Vast.ai
1× RTX 308010 GB	~83 tok/sFaster than you read	$0.082	$0.27	Vast.aiMarketplace · 98.3% reliable	Rent on Vast.ai
1× RTX 30708 GB	~49 tok/sFaster than you read	$0.082	$0.46	Vast.aiMarketplace · 99.1% reliableRunPod $0.13/h	Rent on Vast.ai
1× RTX 4060 Ti 16GB16 GB	~31 tok/sFaster than you read	$0.090	$0.78	Vast.aiMarketplace · 99.4% reliable	Rent on Vast.ai
1× RTX 407012 GB	~55 tok/sFaster than you read	$0.096	$0.48	Vast.aiMarketplace · 99.8% reliable	Rent on Vast.ai
1× RTX 309024 GB	~103 tok/sFaster than you read	$0.12	$0.33	Vast.aiMarketplace · 99.5% reliableRunPod $0.50/h	Rent on Vast.ai
1× RTX 5060 Ti 16GB16 GB	~49 tok/sFaster than you read	$0.12	$0.69	Vast.aiMarketplace · 99.1% reliable	Rent on Vast.ai
1× RTX 3080 Ti12 GB	~100 tok/sFaster than you read	$0.14	$0.37	Vast.aiMarketplace · 99.9% reliable	Rent on Vast.ai

First time renting a GPU? How it works, in four steps

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.
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.
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.
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 Llama-2-7b-hf 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 4,096 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 Llama-2-7b-hf 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.json · pinned 8efe6c9b9365

✓ Architecture read from the published config.json

Retrieved 2026-09-01 at pinned commit 8efe6c9b9365; parameter count from the safetensors index at the same revision.

Architecture: llama
Layers: 32
Hidden size: 4,096
Attention heads: 32
KV heads: 32
Feed-forward width: 11,008
Vocabulary: 32,000
Context ceiling: 4,096

Where the memory goes

Every attention head caches its own keys and values, so the KV cache grows at the full multi-head rate with context.

Open the exact official config.json ↗

More from NousResearch

Hermes-3-Llama-3.1-405B Hermes-3-Llama-3.1-70B Hermes-3-Llama-3.1-8B Hermes-4-405B Hermes-4-70B Meta-Llama-3.1-8B-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":"nousresearch-llama-2-7b-hf","quantization":"Q4_K_M","context":4096,"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="nousresearch-llama-2-7b-hf" data-quantization="Q4_K_M"
  data-hardware="nvidia-geforce-rtx-4090" data-context="4096"></script>

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

Citing this page

LLM Bottleneck. “Llama-2-7b-hf VRAM and hardware requirements.” Architecture from NousResearch/Llama-2-7b-hf at revision 8efe6c9b9365, retrieved 2026-09-01. https://llmbottleneck.com/models/nousresearch-llama-2-7b-hf

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