LLMHOSTING.AI the compute price terminal

Llama 4 Maverick API pricing

9 providers serve Llama 4 Maverick. Llama 4 Maverick is an open-weight model (Llama 4 Community) with 400B total parameters, 17B active per token (mixture-of-experts), up to 1049K context. Prices below are per 1M tokens, cheapest input first (self-hosting math further down).

Refreshed 2026-08-23 · list prices from provider APIs
Input floor $0.050/1M at Lambda Output floor $0.100/1M at Lambda Find your workload winner →
Provider Input $/1M Output $/1M Cache read $/1M Context
Lambda $0.050 $0.100output floor - 131K Use
DeepInfra $0.150 $0.600 - 1049K Use
Groq $0.200 $0.600 - 131K Use
Together AI $0.270 $0.850 - - Use
Novita AI $0.270 $0.850 - 1049K Use
Google Vertex AI $0.350 $1.15 - 1000K Use
SambaNova $0.630 $1.80 - 131K Use
Oracle OCI $0.720 $0.720 - 1049K Use
Azure AI Foundry $1.41 $0.350 - 1000K Use

The cheapest way to run Llama 4 Maverick

On an illustrative 70M-input / 30M-output monthly workload, today's lowest listed cost is roughly $6.50/month, at Lambda. Input-heavy, output-heavy and cache-heavy workloads can produce different winners. Use the live mix calculator below rather than combining floors from two different providers. On output-token cost alone, the API floor remains below our self-hosting estimate of ~$0.350/1M. Details below. Run your own numbers in the breakeven calculator, or read the breakeven math.

Your actual monthly cost

Uses today's list prices

Cache read uses the listed cache rate where available; otherwise it falls back to normal input price. Batch, write-cache, volume and negotiated discounts are excluded.

Or run it yourself: Llama 4 Maverick self-hosting economics

Llama 4 Maverick is open-weight (Llama 4 Community), so the API price above competes with the GPU-hour market. 400B total parameters (MoE, ~17B active per token) needs roughly 551 GB VRAM at FP8 or 276 GB at INT4, KV-cache headroom included.

No single GPU fits Llama 4 Maverick at FP8 (551 GB needed), so it needs a multi-GPU node. Cheapest tracked option: 3× MI300X (192 GB each) at roughly $1.50/hr total (RunPod). Rent 3× MI300X at RunPod → INT4 quantization drops the requirement to 276 GB. Tune it in the calculator.
Breakeven estimate: a well-batched vLLM deployment on 3× MI300X ($1.50/hr) at ~2382 aggregate tok/s and 50% utilization works out to roughly $0.350 per 1M output tokens, versus $0.100 via the cheapest output-token API (Lambda). The API is cheaper unless your utilization or throughput beats these assumptions. Planning estimate: throughput varies with hardware, quantization, batch size and context. Tune it in the calculator.

Related models

ModelCheapest in $/1MCheapest out $/1MProviders
Llama 3.3 70B $0.100 $0.200 15
Llama 3.1 8B $0.020 $0.030 15
Llama 4 Scout $0.050 $0.100 11
Llama 3.1 70B Instruct $0.120 $0.300 10
Llama 3.2 3B $0.020 $0.020 9
Llama-3-70b $0.120 $0.300 7

FAQ

What is the cheapest API for Llama 4 Maverick?

As of 2026-08-23, the lowest input price for Llama 4 Maverick is Lambda at $0.050 per 1M input tokens. The lowest output price is Lambda at $0.100 per 1M output tokens. The cheapest provider for a real workload depends on its input/output mix. The next-lowest input price is DeepInfra at $0.150, a 200% difference.

How much VRAM do you need to self-host Llama 4 Maverick?

Llama 4 Maverick has 400B parameters, so plan for roughly 551 GB of VRAM at FP8 or 276 GB at INT4/AWQ, KV-cache headroom included. That exceeds a single GPU. A typical node is 3× MI300X.

← All models · GPU rental prices · Breakeven calculator · Get this data as JSON