Compare
Llama 4 Scout vs Gemini 2.5 Ultra
Meta vs Google DeepMind. Specs, benchmarks, and real per-task cost — all in one page.
Verdict
Gemini 2.5 Ultra leads on LMSYS ELO (1385 vs 1265). Llama 4 Scout is ~22.7× cheaper on a 3:1 input:output blend. Llama 4 Scout is open-weights, the other is proprietary.
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
Google DeepMind
Gemini 2.5 Ultra2M-token context, native video understanding, and Google’s deepest multimodal stack. The long-context king.
2M context windowNative videoBest multimodal reasoning
Pricing
Input / 1M$0.35$7.00
Output / 1M$0.80$21.00
Context128K2.0M
Max output8K64K
LicenseOpen weightsProprietary
Released2025-12-102026-02-05
Benchmarks
LMSYS ELO1265.01385.0
MMLU Pro77.290.4
HumanEval82.490.2
SWE-Bench—58.3
MATH72.592.0
MMMU—82.1
IFEval84.189.7
Per-task cost
Summarize a 1-hour meeting transcript$0.0057$0.115
Review a 500-line pull request$0.0044$0.098
Answer a customer support ticket$0.0017$0.038
Extract structured data from a resume$0.0026$0.056
Debug a stack trace with context$0.0037$0.084
Per-call cost using published token counts for each task. Real-world prompts vary.