Model by xAI

Grok 4.5

Grok 4.5 is xAI's earlier coding-focused model, and it suits teams that still have it set as the default in Grok Build or Cursor.

Checked 2026-11-05

Grok 4.5 is xAI's earlier coding-focused model, and it suits teams that still have it set as the default in Grok Build or Cursor. The launch post is dated 16 July 2026, though OpenRouter listed the model on 8 July 2026, so treat the launch date as approximate. xAI calls it its strongest model at the time, trained alongside Cursor and serving at about 80 tokens per second. The docs give a 500,000 token context window, text and image input, text output, and reasoning effort from low to xhigh. Price is $2 per million input tokens and $6 per million output tokens under 200,000 prompt tokens, and $4 and $12 at 200,000 or more for the whole request. In the maker's own results it scored 64.7 percent on SWE Bench Pro, behind Opus 4.8 at max effort on 69.2 percent. Free usage in Grok Build and Cursor was time limited, with no end date given. xAI has since released Grok 4.6 and Grok 4.7, and its docs now recommend 4.7. We would not start new work on it. Figures come from xAI's launch post, its docs and OpenRouter. We have not run it ourselves.

Mentioned here

  • Grok 4.6

    Grok 4.6 is xAI's model for long-running coding agents, and it suits teams that already use Grok Build or Cursor.

    • Closed

    xAI, 500K context, $2 in / 1M

  • Grok 4.7

    Grok 4.7 is a closed xAI model, released 2026-09-21, listed on OpenRouter at $2 per million input tokens and $6 output with a 500,000-token context and text, image and PDF input.

    • Closed

    xAI, 500K context, $2 in / 1M

  • Claude Opus 4.8

    Claude Opus 4.8 is the older Opus that many agent setups still run on, and it suits teams that have not yet moved to the 5 series.

    • Closed

    Anthropic, 1M context, $5 in / 1M

  • xAI API

    xAI's Grok API: grok-4.7 $2 in, $6 out with a 500K context. Spend tiers, 30-day audit storage, zero data retention switch, Grok Build and agent setup.

    • Model / inference API
    • OpenAI compatible

  • Grok Build

    Grok Build is xAI's terminal coding agent with plan mode, subagents and MCP. Free to try; Plus $8, Premium $30, Super $100, Ultra $200 a month.

    • Terminal agent
    • MCP

    Subscription

  • Cursor

    Cursor is a desktop AI editor with a terminal CLI and cloud agents that work in their own virtual machines.

    • AI code editor
    • MCP

    Free tier + paid

Questions

What is the context window of Grok 4.5?

500,000 tokens, with up to 450,000 tokens of output.

How much does Grok 4.5 cost?

$2 per million input tokens and $6 per million output tokens, checked 31 October 2026.

Can I run Grok 4.5 on my own hardware?

No. The weights are not published, so you use it through the maker's API or a hosting provider.

Who makes Grok 4.5?

xAI makes it, and released it on 8 July 2026.

Reviews

Loading reviews…

Write a review

Your rating

Up to 1000 characters. Stars alone are enough.

Where you can run it

Compare hosts

  • OpenRouter

    OpenRouter routes more than 600 models through one API key.

    • Model gateway (router)
    • OpenAI compatible

Similar

  • Grok 4.3

    Grok 4.3 is a reasoning model from xAI, now listed by OpenRouter under the name SpaceXAI, that suits people who want a one million token context at a low output price.

    • Closed

    xAI, 1M context, $1.25 in / 1M