Compare
GPT-5 Turbo vs Llama 4 Behemoth
OpenAI vs Meta. Specs, benchmarks, and real per-task cost — all in one page.
Verdict
GPT-5 Turbo leads on LMSYS ELO (1398 vs 1342). Llama 4 Behemoth is ~2.7× cheaper on a 3:1 input:output blend. Llama 4 Behemoth is open-weights, the other is proprietary.
OpenAI
GPT-5 TurboOpenAI’s flagship unified model. Handles text, vision, and audio natively. The generalist benchmark champion.
Native multimodalStrong math and scienceHuge ecosystem
Meta
Llama 4 BehemothMeta’s open-weights flagship. 405B params, fully open license, runs on every major inference provider.
Best open-weights modelRuns anywhereTransparent training
Pricing
Input / 1M$6.00$2.50
Output / 1M$24.00$8.00
Context400K256K
Max output32K16K
LicenseProprietaryOpen weights
Released2026-01-212025-12-10
Benchmarks
LMSYS ELO1398.01342.0
MMLU Pro91.287.1
HumanEval93.089.3
SWE-Bench65.852.4
MATH91.581.5
GPQA68.156.2
MMMU78.8—
IFEval88.988.0
Per-task cost
Summarize a 1-hour meeting transcript$0.102$0.041
Review a 500-line pull request$0.096$0.036
Answer a customer support ticket$0.037$0.014
Extract structured data from a resume$0.054$0.021
Debug a stack trace with context$0.084$0.031
Per-call cost using published token counts for each task. Real-world prompts vary.