OpenGPTEvidence-based model decisionsModel families
Llama model family
The Llama family currently contains 3 provider-documented model IDs in the OpenGPT catalog. Each version keeps its own identity and lifecycle.
- Official model IDs
- 3
- Model providers
- Meta
- Mapped public signals
- 3
- Identity reviewed
Evidence scope
This page is an index of existing public evidence and provider-documented identities. Inclusion does not imply an OpenGPT endorsement or a universal rank. Meta · Llama.
How OpenGPT handles this evidence
Official identities come from provider documentation. Public ranking signals are attached only where the source label and exact catalog version have a reviewed mapping.
Comparison boundaries
- Models in one provider or family are not interchangeable; mode, version, lifecycle, tools, and service tier may change results.
- Absence from a public ranking means no comparable published signal was mapped, not a score of zero.
- Provider documentation establishes identity, not independent quality or suitability.
Official model IDs
The Llama family currently contains 3 provider-documented model IDs in the OpenGPT catalog. Each version keeps its own identity and lifecycle.
- Llama 3.3 70B Instructmeta-llama/Llama-3.3-70B-InstructStableIdentity reviewed: 2026-07-24Mapped public signals: 1
- Llama 4 Maverick 17B-128Emeta-llama/Llama-4-Maverick-17B-128E-InstructStableIdentity reviewed: 2026-07-24Mapped public signals: 1
- Llama 4 Scout 17B-16Emeta-llama/Llama-4-Scout-17B-16E-InstructStableIdentity reviewed: 2026-07-24Mapped public signals: 1