Meta cloud models
17 models tracked via Artificial Analysis. Compare cloud performance, then find local GGUF versions in the GraySoft model catalog.
| Model | Intelligence | Speed (tok/s) |
|---|---|---|
| Muse Spark | 52.2 | 0 |
| Llama 4 Maverick | 18.4 | 95.349 |
| Llama 3.1 Instruct 405B | 17.4 | 61.926 |
| Llama 3.3 Instruct 70B | 14.5 | 94.195 |
| Llama 4 Scout | 13.5 | 96.052 |
| Llama 3.1 Instruct 70B | 12.5 | 35.361 |
| Llama 3.2 Instruct 90B (Vision) | 11.9 | 47.103 |
| Llama 3.1 Instruct 8B | 11.8 | 197.755 |
| Llama 3.2 Instruct 3B | 9.7 | 52.169 |
| Llama 2 Chat 7B | 9.7 | 121.664 |
| Llama 3 Instruct 70B | 8.9 | 43.958 |
| Llama 3.2 Instruct 11B (Vision) | 8.7 | 86.713 |
| Llama 2 Chat 70B | 8.4 | 0 |
| Llama 2 Chat 13B | 8.4 | 0 |
| Llama 65B | 7.4 | 0 |
| Llama 3 Instruct 8B | 6.4 | 80.036 |
| Llama 3.2 Instruct 1B | 6.3 | 87.116 |
Run models locally with guIDE
Download guIDE — the AI-native code editor with local LLM inference and 69 built-in tools.