Grok 4 — Specifications
| Sviluppatore | xAI |
|---|---|
| Tipo | Large language model (reasoning) |
| Modalità | Text + image input, text output |
| Parametri | Non divulgato |
| Finestra contestuale | 256,000 tokens |
| Output massimo | Non divulgato |
| Licenza | Proprietario |
| Pesi aperti | ❌ No |
| Data di rilascio | July 9, 2025 |
| Costo per input | $3 /1M |
| Costo per output | $15 /1M |
| Provider API | xAI API, OpenRouter, Vercel AI Gateway |
Grok 4 is xAI’s flagship reasoning model, launched on 9 July 2025 as the successor to Grok 3. It pairs native chain-of-thought reasoning with multimodal understanding, accepting both text and image input and returning text. The model ships with a 256,000-token context window and was, at release, xAI’s most capable system, posting strong results on reasoning and tool-use benchmarks such as Humanity’s Last Exam and ARC-AGI, aided by real-time access to X and the web.
Access is entirely proprietary: there are no open weights, and the model runs through the xAI API, the Grok apps, and the SuperGrok subscription. API pricing sits at $3 per million input tokens and $15 per million output tokens, placing it in the same bracket as frontier models from OpenAI and Anthropic rather than the budget tier.
Verdetto: Grok 4 is a genuine frontier-class reasoner with strong benchmarks and live search, but it is not cheap and has since been joined by faster, lower-cost variants (Grok 4 Fast, 4.20, 4.3). Choose it for hard reasoning inside xAI’s ecosystem; reach for a newer variant when latency or price matters more than peak capability.
