Compare
GPT-5 Mini vs Llama 4 Scout
OpenAI vs Meta. Specs, benchmarks, and real per-task cost — all in one page.
Verdict
GPT-5 Mini leads on LMSYS ELO (1288 vs 1265). Llama 4 Scout is ~1.9× cheaper on a 3:1 input:output blend. Llama 4 Scout is open-weights, the other is proprietary.
OpenAI
GPT-5 MiniGPT-5’s smaller, cheaper sibling. Optimized for chat and lightweight agent tasks.
Very low latencyCheap enough for mass consumer apps
Meta
Llama 4 ScoutThe 70B workhorse open-weight. Fits on a single H100, beloved by self-hosters.
Single-GPU deployableDirt cheap on aggregatorsStrong fine-tuning base
Pricing
Input / 1M$0.50$0.35
Output / 1M$2.00$0.80
Context256K128K
Max output16K8K
LicenseProprietaryOpen weights
Released2026-01-212025-12-10
Benchmarks
LMSYS ELO1288.01265.0
MMLU Pro78.577.2
HumanEval85.582.4
MATH79.872.5
IFEval85.484.1
Per-task cost
Summarize a 1-hour meeting transcript$0.0085$0.0057
Review a 500-line pull request$0.0080$0.0044
Answer a customer support ticket$0.0031$0.0017
Extract structured data from a resume$0.0045$0.0026
Debug a stack trace with context$0.0070$0.0037
Per-call cost using published token counts for each task. Real-world prompts vary.