Compare

Llama 4 Scout vs Claude Haiku 4.5

Meta vs Anthropic. Specs, benchmarks, and real per-task cost — all in one page.

Verdict

Claude Haiku 4.5 leads on LMSYS ELO (1305 vs 1265). Llama 4 Scout is ~3.5× cheaper on a 3:1 input:output blend. Llama 4 Scout is open-weights, the other is proprietary.

Meta

Llama 4 Scout

The 70B workhorse open-weight. Fits on a single H100, beloved by self-hosters.

Single-GPU deployableDirt cheap on aggregatorsStrong fine-tuning base

Anthropic

Claude Haiku 4.5

Fast, cheap, and smart enough for most routing and extraction tasks. Great base model for subagents.

Sub-second latency10x cheaper than OpusStrong structured output

Pricing

Input / 1M$0.35$0.80

Output / 1M$0.80$4.00

Context128K200K

Max output8K16K

LicenseOpen weightsProprietary

Released2025-12-102025-10-01

Benchmarks

LMSYS ELO1265.01305.0

MMLU Pro77.281.4

HumanEval82.487.0

SWE-Bench—49.2

MATH72.576.8

IFEval84.187.3

Per-task cost

Summarize a 1-hour meeting transcript$0.0057$0.014

Review a 500-line pull request$0.0044$0.014

Answer a customer support ticket$0.0017$0.0056

Extract structured data from a resume$0.0026$0.0080

Debug a stack trace with context$0.0037$0.013

Per-call cost using published token counts for each task. Real-world prompts vary.

Tool

Interactive compare →

Add up to four models, tweak tasks live.