Grok model for agentic tool use, reasoning, coding, and live assistance
Details
| Field | Value |
|---|---|
| id | grok-4-20-non-reasoning |
| family | grok |
| modality | text+image+pdf->text |
| context_length | 2000000 |
| max_output_tokens | 30000 |
| release_date | 2026-03-09 |
| is_open_weights | no |
| supports_tool_call | yes |
| supports_reasoning | no |
| supports_structured_output | yes |
| supports_attachment | yes |
| input_modalities | text, image, pdf |
| output_modalities | text |