Skip to content
Tokenly.
Demo fixture — not an offerDemo fixture

Llama 3.1 70B Instruct

meta-llama/Llama-3.1-70B-Instruct

A 70-billion-parameter open-weight model positioned by Meta for general instruction-following and assistant workloads. Supports a 128K-token context window and tool calling.

Llama 3.1 70B Instruct
Model IDmeta-llama/Llama-3.1-70B-Instruct
DeveloperMeta
Serving providerPending confirmation
Capabilitieschatreasoningfunction-callinglong-context
Modalitiestext
Context window128K tokens
Pricing
Input
Not yet published
Output
Not yet published
LimitationsText-only input. Token limits and hosted behaviour vary by serving provider.
SourceModel card ↗
Last verified18 Nov 2026