Skip to main content
Meta·current

Llama 3.1 405B

Meta's largest open-weight model — GPT-4-class intelligence, fully open

ProviderMeta
Released2024-07-23
StatusCurrent
Context window131K tokens
PricingOpen
Capabilities
textcodefunction-callingreasoning
Hugging FaceView on HF ↗

Benchmarks

MMLU (knowledge)88.6
MATH (mathematics)73.5

As of 2024-12-01. Source

Loading…

Compare with similar models

Llama 3.1 405BDeepSeek V3DeepSeek R1Llama 3.3 70B
Context131K66K66K131K
MMLU88.688.590.886.0
HumanEval91.6
MATH73.590.297.377.0
GPQA71.5
PricingOpenOpenOpenOpen
Input $/M$0.27$0.55

Compare more models →