Llama 4 Maverick
Open weightsApr 5, 2025
- Context
- 1M tokens
- Max output
- —
- Input / 1M
- —
- Output / 1M
- —
0 mapped APIs
Llama 4 is Meta's open-weight multimodal mixture-of-experts model family, represented in iilib by Scout and Maverick.
Llama 4 Scout and Maverick combine text and image input with multilingual text output. Scout emphasizes extreme long context while Maverick uses a larger expert pool for general assistant and multimodal workloads.
Current indexed models and their documented specifications.
| Model | Released | Context | Max output | Input | Output | API |
|---|---|---|---|---|---|---|
| Llama 4 Maverick Open weights | Apr 5, 2025 | 1M tokens | — | — | — | — |
| Llama 4 Scout Open weights | Apr 5, 2025 | 10M tokens | — | — | — | — |
May we use Google Analytics to understand which research pages and tools are useful? Analytics is off unless you agree. We never send your calculator inputs or search text as custom events.Privacy details.