Grok 1 Models by xAI
xAI releases · 15 in the last 12 monthsChangelog →
4 models2023–2024
Last refreshed 2026-05-22. Next refresh: weekly.
Details
ResearcherxAI
LicenseApache 2.0OSI-approved
Commercial useCommercial use: permitted
Models4
Released2023–2024
Capabilities
ReasoningAll models
About
Grok 1 is xAI’s large open-weight language model: a 314 billion parameter Mixture-of-Experts model trained from scratch by xAI. The open release shares the raw base-model checkpoint from the Grok-1 pre-training phase—not fine-tuned for dialogue—with network architecture details, JAX example code on GitHub, and weights on Hugging Face. The published design uses a mixture of 8 experts (2 active per token), 64 layers, and an 8,192-token context.
Decision facts
Current Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
4 in view
Grok-1Current
Use when the workload needs 314B parameters and reasoning.
2023-11314B parametersreasoning
Grok-0Current
Use when the workload needs 33B parameters and reasoning.
2023-0833B parametersreasoning
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Grok-1.5V | Use when the workload needs reasoning. | 2024-04 | reasoning | Current |
| Grok-1.5 | Use when the workload needs reasoning. | 2024-03 | reasoning | Current |
| Grok-1 | Use when the workload needs 314B parameters and reasoning. | 2023-11 | 314B parametersreasoning | Current |
| Grok-0 | Use when the workload needs 33B parameters and reasoning. | 2023-08 | 33B parametersreasoning | Current |
