xAI's new flagship — faster, more token-efficient, and lower cost than the prior generation, ranked near the top of the Artificial Analysis Intelligence Index.
Grok 4.5 keeps the flagship rate at $2 / 1M input and $6 / 1M output, with cached input at $0.50 / 1M. A higher-context surcharge applies to requests above 200K tokens.
xAI revealed Grok 4.5 in private beta on June 28, 2026, then launched it for developers on July 8 and to the public on July 9. xAI calls it an Opus-class model — faster, more token-efficient, and lower cost than the Grok 4.20 flagship it replaces, at the same $2 / $6 per-1M headline rate. It reportedly ranks near the top of the Artificial Analysis Intelligence Index, a strong showing for a model priced well below Claude Opus 4.8.
The model ships a 500K-token context window with configurable low/medium/high reasoning effort, function calling, web and X search, and code execution. Prompt caching cuts cached input to $0.50 / 1M — a 75% discount — which matters for agents that reuse large system prompts.
Grok 4.5 is built for coding and agentic work: tool-use loops, code execution, and real-time retrieval from the web and X. Its token efficiency and low output price make it attractive for high-volume agent runs, while native X search keeps its edge on current events and social context. Availability spans the xAI API console, Grok Build, Cursor, OpenRouter, Vercel, Cloudflare, and Microsoft Office add-ins.
| Model | Input / 1M | Output / 1M | Context |
|---|---|---|---|
| Grok 4.5 | $2 | $6 | 500K |
| Grok 4.1 Fast | $0.20 | $0.50 | 2M |
| Claude Opus 4.8 | $5 | $25 | 1M |
| GPT-5.6 Sol | $5 | $30 | 1M |
| Gemini 3.1 Pro | $2 | $12 | 1M |
Grok 4.5 undercuts the other flagships sharply on output price — half of Gemini 3.1 Pro and a fifth of GPT-5.6 Sol — while claiming Opus-class quality. For the longest contexts and cheapest tokens, Grok 4.1 Fast remains xAI's speed-and-scale option.