Model comparison
Llama 4 Maverick vs Grok 3
Meta
Llama 4 Maverick
The bigger Llama 4 — frontier quality you can self-host.
xAI
Grok 3
xAI's bet on inference compute, wired into X for fresh context.
Specs
| Metric | Llama 4 Maverick | Grok 3 |
|---|---|---|
| Context window | 1.0M tokens↑ | 131K tokens |
| Input $/1M tokens | $0.150↑ | $3.00 |
| Output $/1M tokens | $0.600↑ | $15.00 |
| Modalities | Text · Image | Text |
| Open weights | Yes | No |
| Released | Apr 2025 | — |
Capability differences
| Capability | Llama 4 Maverick | Grok 3 |
|---|---|---|
| Extended thinking | No | Yes |
| Open weights | Yes | No |
Key differences
- —Llama 4 Maverick offers 1.0M context; Grok 3 offers 131K.
- —Input pricing: Llama 4 Maverick at $0.150/1M vs Grok 3 at $3.00/1M.
- —Llama 4 Maverick is open weights; Grok 3 is proprietary.