Meta
Llama 4 Scout
active open-weight
Smaller MoE Llama 4 model. 109B active params, massive 10M context window.
Context 10M
Max output 64K
Pricing /1M tok $0.10 / $0.40 • OpenRouter
Released Apr 5, 2025
// provider_directory
Meta
All tracked Meta models — pricing, context windows, lifecycle state, and release history in one place.
10 tracked models
5 live or preview
10 with listed pricing
Apr 5, 2025 latest release
// model_index
All Meta models
Meta
Llama 4 Maverick
active open-weight
Meta's MoE model with 400B active parameters. 1M context, vision support. Strong performance at low cost.
Context 1M
Max output 64K
Pricing /1M tok $0.15 / $0.60 • OpenRouter
Released Apr 5, 2025
Lifecycle
← llama-3-3-70b
Meta
Llama 3.3 70B
active open-weight
Strong 70B dense model. 128K context. Best Llama 3.x text-only model.
Context 128K
Max output 8K
Pricing /1M tok $0.12 / $0.30 • OpenRouter
Released Dec 6, 2024
Lifecycle
← llama-3-1-70b → llama-4-maverick
Meta
Llama 3.2 11B Vision
active open-weight
11B multimodal model. 128K context. Good efficiency for vision tasks.
Context 128K
Max output 8K
Pricing /1M tok $0.05 / $0.10 • OpenRouter
Released Sep 25, 2024
Meta
Llama 3.2 90B Vision
active open-weight
90B multimodal model. 128K context. Best Llama 3.x vision model.
Context 128K
Max output 8K
Pricing /1M tok $0.18 / $0.45 • OpenRouter
Released Sep 25, 2024
Lifecycle
→ llama-4-maverick
Meta
Llama 3.1 405B
legacy open-weight
405B dense model. Was the largest open-weight model at release. Superseded by Llama 4.
Context 128K
Max output 8K
Pricing /1M tok $0.55 / $0.80 • OpenRouter
Released Jul 23, 2024
Lifecycle
→ llama-4-maverick
Meta
Llama 3.1 70B
legacy open-weight
70B dense model. Standard workhorse before Llama 3.3.
Context 128K
Max output 8K
Pricing /1M tok $0.12 / $0.30 • OpenRouter
Released Jul 23, 2024
Lifecycle
← llama-3-70b → llama-3-3-70b
Meta
Llama 3.1 8B
legacy open-weight
8B dense model. Smallest in the Llama 3.1 family. Good for local inference.
Context 128K
Max output 8K
Pricing /1M tok $0.02 / $0.05 • OpenRouter
Released Jul 23, 2024
Lifecycle
← llama-3-8b
Meta
Llama 3 70B
legacy open-weight
Original Llama 3 70B. Only 8K context. Superseded by 3.1 with 128K.
Context 8K
Max output 4K
Pricing /1M tok $0.12 / $0.30 • OpenRouter
Released Apr 18, 2024
Lifecycle
→ llama-3-1-70b
Meta
Llama 3 8B
legacy open-weight
Original Llama 3 8B. Only 8K context. Superseded by 3.1 with 128K.
Context 8K
Max output 4K
Pricing /1M tok $0.02 / $0.05 • OpenRouter
Released Apr 18, 2024
Lifecycle
→ llama-3-1-8b
// lineage_chains
Model lineage
Predecessor → successor chains tracked for Meta models.
Llama 3.2 90B Vision active Sep 25, 2024
→
Llama 4 Maverick active Apr 5, 2025
Llama 3.1 405B legacy Jul 23, 2024
→
Llama 4 Maverick active Apr 5, 2025
Llama 3 70B legacy Apr 18, 2024
→
Llama 3.1 70B legacy Jul 23, 2024
→
Llama 3.3 70B active Dec 6, 2024
→
Llama 4 Maverick active Apr 5, 2025
Llama 3 8B legacy Apr 18, 2024
→
Llama 3.1 8B legacy Jul 23, 2024
// release_timeline
Meta timeline
Llama 4 Maverick released
Meta's MoE model with 400B active parameters. 1M context, vision support. Strong performance at low cost.
Llama 4 Scout released
Smaller MoE Llama 4 model. 109B active params, massive 10M context window.
Llama 3.3 70B released
Strong 70B dense model. 128K context. Best Llama 3.x text-only model.
Llama 3.2 11B Vision released
11B multimodal model. 128K context. Good efficiency for vision tasks.
Llama 3.2 90B Vision released
90B multimodal model. 128K context. Best Llama 3.x vision model.
Llama 3.1 405B released
405B dense model. Was the largest open-weight model at release. Superseded by Llama 4.
Llama 3.1 70B released
70B dense model. Standard workhorse before Llama 3.3.
Llama 3.1 8B released
8B dense model. Smallest in the Llama 3.1 family. Good for local inference.
Llama 3 70B released
Original Llama 3 70B. Only 8K context. Superseded by 3.1 with 128K.
Llama 3 8B released
Original Llama 3 8B. Only 8K context. Superseded by 3.1 with 128K.