Model comparison
Grok 4.5 vs Grok 3
Capability radar, value bars, and side-by-side posture.
Grok 4.5
Not scored
value score
Grok 3
89
value score
Grok 4.5 ctx
—
Grok 3 ctx
131K
Capability radar
Value · context · multimodal · openness · speed posture
Value head-to-head
- Grok 389
Attribute tape
| Field | Grok 4.5 | Grok 3 |
|---|---|---|
| Developer | xAI | xAI |
| Context | — | 131K |
| Modalities | Text, Code | |
| Openness | — | Proprietary Model |
| Speed | — | — |
| Price | — | — |
| Value score | — | 89 |
| Summary | Grok 4.5 is xAI's frontier mixture-of-experts model optimized for coding, agentic tasks, and knowledge work. Launched in July 2026, it is available via the xAI developer API, Grok Build, and Cursor. | xAI frontier reasoning model. |