What is Llama 4 Maverick good for?

Text and image work over a very long context. Open weights under the Llama 4 Community License, on a multi-GPU node.

When is Llama 4 Maverick not a good fit?

Self-hosting on small hardware: memory follows the full 400B parameters, not the 17B active. No video input or structured output, and its knowledge is older than the current open models'.

What is the cheapest way to run Llama 4 Maverick?

Hosted, unless you push serious volume. DigitalOcean charges $0.20 in / $0.70 out per 1M tokens. The cheapest rental that fits is 8x RTX A6000 at $3,168 a month, which costs the same as about 11B tokens a month on that API.

Can I self-host Llama 4 Maverick?

Yes. The weights are released under the Llama 4 Community License. At 4-bit it needs about 228 GB of GPU memory, which starts at roughly $3,168 a month on the cheapest rental that fits.

Open weights · Llama 4 Community License

Llama 4 Maverick

Meta's first MoE Llama generation: a 400B mixture-of-experts model with 17B active across 128 experts that fits on a single H100 node.

Jump to: Specs Pricing Self-hosting Similar models FAQ More from Meta

Visit website Compare vs other LLMs →

Cheapest via DigitalOcean

Input: $0.20 /1M tokens
Output: $0.70 /1M tokens

5 providers

Key Specifications

Context window: 1M tokens
Knowledge cutoff: Aug 2024
Parameters: 400B, 17B active
Inputs: Text, image

Capabilities Show details: Function calling Connect to external tools, APIs, and systems.

Hosted API pricing

Input tokens M Output tokens M

Provider	Input /1M tokens	Output /1M tokens	Cost at 10M in + 2M out
DigitalOcean Cheapest	$0.20	$0.70	$3.40	View
Replicate	$0.25	$0.95	$4.40	View
Azure	$0.25	$1.00	$4.50	View
Novita	$0.27	$0.85	$4.40	View
Meta Creator	Open weights, no hosted price			View

Heads up: Base-tier, on-demand rates per 1M tokens; cached, batch and long-context tiers excluded. A provider may serve a shorter context or a quantized build than the creator's release. Verify before provisioning. More on how we price.

Estimated cost to self-host

Llama 4 Maverick needs about 228 GB of GPU memory at 4-bit with 32K context. The cheapest rental that fits is 8× RTX A6000 at about $3,168 a month.

That costs the same as roughly 11B tokens a month on DigitalOcean's API. Self-hosting is more expensive below that volume.

Single node ≤ 8

Precision	Memory	Cheapest fit (32K context)(Full 1M context)	Cost /mo	Break-even vs API
4-bitINT4 / FP4	228 GB320 GB	8× RTX A6000	$3,168$4,896	11B tokens /mo17B tokens /mo
8-bitFP8 / INT8	408 GB500 GB	8× A100	$10,138$11,362	36B tokens /mo40B tokens /mo
16-bitFP16 / BF16	808 GB999 GB	4× B300	$22,723$22,723	80B tokens /mo80B tokens /mo

Estimates based on median on-demand rates for Nvidia GPUs. Memory is weights plus KV cache for one request, using FP8 KV cache where supported and FP16/BF16 otherwise. Break-even assumes a 5:1 input-to-output ratio. No guarantee of runtime support, usable performance, or that a matching quantized build exists. How we estimate costs.

Similarly priced models

The models nearest Llama 4 Maverick by blended rate, each at its own cheapest provider.

Model	Blended /1M	Input /1M	Output /1M	Context	Cutoff	vs Llama 4 Maverick
Mistral Small 4 Mistral	$0.225	$0.15	$0.60	256K		−21%
Qwen3.8-27B Alibaba Cloud	$0.225	$0.17	$0.50	262K		−21%
Qwen3-Coder-Next Alibaba Cloud	$0.23	$0.08	$0.98	262K		−19%
Llama 4 Scout Meta	$0.24	$0.16	$0.64	10M	Aug 2024	−15%
Qwen3.5-122B-A10B Alibaba Cloud	$0.2533	$0.12	$0.92	262K		−11%
Llama 4 Maverick This model Meta	$0.2833	$0.20	$0.70	1M	Aug 2024
Qwen3-VL-30B-A3B Instruct Alibaba Cloud	$0.2833	$0.20	$0.70	262K		0%
Muse Glimmer 30B Meta	$0.30	$0.20	$0.80	131K	Jan 2026	+6%
Qwen2.5-7B Instruct Alibaba Cloud	$0.30	$0.30	$0.30	33K		+6%
Qwen3-235B-A22B FP8 Alibaba Cloud	$0.30	$0.20	$0.80	33K		+6%
Qwen3.5-397B-A17B Alibaba Cloud	$0.3133	$0.17	$1.03	262K		+11%

Prices are USD per 1M tokens at each model's cheapest listed provider. Blended is the cost of 10M input plus 2M output tokens, spread over the 12M.

Thunder Compute

RTX A6000 from $0.35, A100 from $1.09, per GPU per hour

Billed per minute, no commitment, no minimums

One click into VS Code or Cursor, no SSH setup

See live pricing

Frequently Asked Questions

What is Llama 4 Maverick good for?: Text and image work over a very long context. Open weights under the Llama 4 Community License, on a multi-GPU node.
When is Llama 4 Maverick not a good fit?: Self-hosting on small hardware: memory follows the full 400B parameters, not the 17B active. No video input or structured output, and its knowledge is older than the current open models'.
What is the cheapest way to run Llama 4 Maverick?: Hosted, unless you push serious volume. DigitalOcean charges $0.20 in / $0.70 out per 1M tokens. The cheapest rental that fits is 8x RTX A6000 at $3,168 a month, which costs the same as about 11B tokens a month on that API.
Can I self-host Llama 4 Maverick?: Yes. The weights are released under the Llama 4 Community License. At 4-bit it needs about 228 GB of GPU memory, which starts at roughly $3,168 a month on the cheapest rental that fits.

More from Meta

Model	Context	Cutoff	Input /1M	Output /1M
Muse Glimmer 30B Open weights	131K	Jan 2026	$0.20	$0.80
Llama 4 Scout Open weights	10M	Aug 2024	$0.16	$0.64
Llama 3.3 70B Open weights	128K	Dec 2023	$0.13	$0.40
Llama 3.2 11B Vision Open weights	128K	Dec 2023	$0.05	$0.68
Llama 3.2 3B Open weights	128K	Dec 2023	$0.05	$0.34
Llama 3 8B Open weights	8K	Mar 2023	$0.05	$0.25
Llama 3 70B Open weights	8K	Dec 2023	$0.65	$2.75
Llama 3.2 1B Open weights	128K	Dec 2023	$0.03	$0.20