ModelPriceWatch.com
Last scan 2026-08-28 Models tracked 237 Providers 31 Cheapest paid Granite 4.0 H Micro $0.017/Mtok in Every price links to its source

Llama 4 Scout

by Meta · 17B (16 experts) parameters

Current open weights open weights cheap tier
Email me when Llama 4 Scout gets cheaper

Today's price · per 1M tokens

Input

$0.110

per 1M tokens

Output

$0.340

per 1M tokens

Blended

$0.168

blended $/1M — 3:1 weighted input:output

Source: official Meta pricing · read Jun 25, 2026 MODELPRICEWATCH.COM · 2026-08-28

Price receipt

No confirmed receipt. We hold dated captures of Meta's pricing page, but none of them tied Llama 4 Scout to the price above on that page, so we do not claim this number is sourced. Check the provider's page directly before relying on it.

Available on 2 hosts

cheapest blended first · $ per 1M tokens

Llama 4 Scout is sold by 2 providers. Prices are per 1M tokens (blended = (3×input + 1×output) ÷ 4). The first-party row is the model maker; “vs first-party” shows each host’s blended price relative to it.

Llama 4 Scout price on every host still selling it, per 1M tokens, cheapest blended first
Host Input Output Blended vs first-party
DeepInfra details → $0.100 $0.300 $0.150 -10%
Meta first-party $0.110 $0.340 $0.168 —
Each host links to its provider page and official pricing. Prices refresh twice daily. MODELPRICEWATCH.COM · 2026-08-28

Overview

Meta's open-weight Llama 4 Scout — a natively multimodal MoE (17B active / 109B total, 16 experts) with an industry-leading 10M-token context, the longest of any open-weight model. Handles text, images, and video; fits on a single H100 with Int4. $0.11/$0.34 per 1M tokens.

Run it yourself

Deploy this open model on rented GPUs

Open weights mean you can self-host instead of paying per-token API prices. These platforms let you serve it on demand — often cheaper at scale.

RunPod →
Rent GPUs by the second and serve any open-weight model yourself.
Novita AI →
Serverless open-model inference billed per token — no GPU ops.
Vast.ai →
GPU marketplace — rent spot/on-demand GPUs cheaply and serve any open-weight model.

Partner links — we may earn a commission if you sign up, at no cost to you. We only list platforms we'd recommend regardless.

Capabilities

struck through = not supported
Input 2/5
Text ✓ Image ✓ Audio Video PDF
Output 1/5
Text ✓ Image Audio Video Embedding
Features 3/9
Prompt caching Reasoning Coding Fast inference Long context ✓ Open weights ✓ Multimodal ✓ Web search Realtime

Benchmark performance

accuracy % · higher is better
Avg benchmark score46.8%
Perf per $/Mtok279.4
GPQA Diamond
57.2%
SWE-Bench Verified
28%
LiveCodeBench
32.8%
Humanity's Last Exam
12%
ARC-AGI 2
15%
AIME 2025
45%
MMMLU
75%
BFCL
58%
HumanEval
80%
MATH 500
65%
Independent composite scores
  • AA Intelligence Index: 10.3 (Artificial Analysis composite across reasoning, knowledge and coding evals)
  • AA Agentic Index: 1.1 (Tool use, planning, autonomy)
  • AA-Omniscience Index: -52.1 (−100–100)
How it stacks up
  • Ranks #63 of 73 benchmarked models by average score
  • Ranks #138 of 170 comparably-measured models by percentile score, across 13 independent measurements
  • Ranks #22 of 170 comparably-tested models by normalized performance per dollar
  • Strongest at HumanEval — 80%, #25 of 49

2600 tokens/sec output 0.33s latency to first token (TTFT)

Source: Vellum LLM Leaderboard · updated Aug 28, 2026 · See full rankings →

Specifications

Provider
Meta
Context window
10M tokens
Modality
text, image
Parameters
17B (16 experts)
Open source
Yes — open weights available
Released
Apr 6, 2025
Status
Current
Last updated
Jun 25, 2026
Tags
multimodalopen-weightsmoe

Availability verified: Jul 25, 2026 — listed on Meta's own page

Official pricing page