CrossModel
Grok 4.6
x-ai/grok-4.6
Back to model catalog

xAI · Model guide

Grok 4.6

Compare
x-ai/grok-4.6

xAI's frontier model for long-running agents, coding and knowledge work. It builds on Grok 4.5 with better persistence across many-step tasks and stronger first passes on interactive and visual projects, and pairs a 500K context window with four reasoning efforts up to xhigh.

Modalities
TextImageText
Context
500K
Max output
500K
Price / 1M tokens
$ 2.00in$ 6.00out
StreamingTool callingJSON outputImage inputThinking mode

What this model is good at

What the vendor positions it for, and which of our models to reach for instead.

VendorSpaceXAI's frontier model built for coding, agentic tasks and knowledge work.source

Best for

  • Long-running coding agents · The launch evaluations centre on agentic engineering — Terminal-Bench 3.0, CursorBench, DeepSWE and FrontierCode.source

  • Knowledge work · Official results cover GDPval-AA v2 and AA-Briefcase, both knowledge-work evaluations.source

  • Interactive and visual builds · xAI reports stronger first passes on visual and interactive projects than Grok 4.5, establishing structure and visual language in one pass.(vendor claim)source

Neighbouring models
  • x-ai/grok-4.3longer context —Use Grok 4.3 for a 1M-token context window and an optional non-reasoning mode.
  • x-ai/grok-4.5predecessor —Use Grok 4.5 to stay on the previous generation of the same family.
  • x-ai/grok-build-0.1sibling —Use Grok Build 0.1 for fast, coding-specialized agent loops.
Compare them side by side →

Pricing and billing

Tiered pricing: the rate changes once the input passes the threshold.

Price / 1M tokens by input sizeInput < 200.0KInput ≥ 200.0K
Input$ 2.00$ 4.00
Output$ 6.00$ 12.00
Cache read$ 0.50$ 1.00
Cache write$ 2.00$ 4.00
Estimate a request

At this input size you are on the Input < 200.0K tier.

$ 0.0320
Estimated cost per request
10.0K × $ 2.00
+ 2.0K × $ 6.00
≈ $ 32.00 per 1,000 requests

Capabilities and limits

What CrossModel guarantees across every route this model can take right now.

Context window
500.0K tokens
Max output
500.0K tokens
Input / output modalities
Text + Image → Text
Streaming
Supported
Tool calling
Supported
Structured output (JSON)
Supported
Image input
Supported
Thinking mode
Supported· always on
Reasoning effort
low · medium · high · xhigh
Available endpoints
/v1/chat/completions · /v1/responses · /v1/messages
First-party sources:Specsdocs.x.ai 1x.ai 2Capabilitiesdocs.x.ai 1docs.x.ai 2

Published benchmarks

Scores the vendor reported, with the evaluation setup each one came from.

Agentic coding
  • CursorBench v3.2
    score·SpaceXAI · 2026-08-12effort highsetup not fully disclosed
    69.9%
  • FrontierCode v1.1 (Extended)
    score·SpaceXAI · 2026-08-12effort highsetup not fully disclosed

    Extended 与 Main 是不同切分，分数不可互比。

    61.3%
  • Terminal-Bench 3.0
    pass_rate·SpaceXAI · 2026-08-12effort highsetup not fully disclosed

    3.0 比 2.x 难得多，绝对分与 Terminal-Bench 2.0 / 2.1 不可比。

    26%
Agentic work
  • APEX-Agents
    score·SpaceXAI · 2026-08-12effort highsetup not fully disclosed
    57.5%
Coding
  • APEX-SWE
    score·SpaceXAI · 2026-08-12effort highsetup not fully disclosed
    56.4%
  • DeepSWE 1.1
    resolved_rate·SpaceXAI · 2026-08-12effort highsetup not fully disclosed

    发布页未披露 harness；同表把 Grok 4.5 记作 54%，与 4.5 发布页的 53% 有出入，两处分别按各自来源保留。

    65.9%
Knowledge work
  • AA-Briefcase
    score·SpaceXAI · 2026-08-12effort highsetup not fully disclosed

    Artificial Analysis 的评分制指标，不是百分比。

    1577
  • GDPval-AA v2
    elo·SpaceXAI · 2026-08-12effort highsetup not fully disclosed

    Artificial Analysis 的知识工作 Elo 评分。

    1753
  • Harvey LAB (Vals)
    score·SpaceXAI · 2026-08-12effort highsetup not fully disclosed

    由 Vals AI 运行的法律领域评测；绝对分整体偏低，仅在同表模型间比较有意义。

    15.8%
Reasoning
  • AA Intelligence Index
    score·SpaceXAI · 2026-08-12effort highsetup not fully disclosed

    Artificial Analysis 的复合指数，由九项评测汇总，不是百分比。xAI 在发布页转载该分数。

    61

Vendor-reported numbers, not CrossModel measurements. Scores are only comparable when the benchmark version, metric and evaluation setup match, so nothing here is averaged or ranked.

Use it in your tools

Point the base URL at CrossModel and paste this model ID — every tool below has a setup guide.

Claude Code
Anthropic's official command-line coding assistant
Cursor
AI-first code editor
Cline
Open-source AI coding agent for VS Code
Codex
OpenAI's coding agent for ChatGPT desktop and the terminal
OpenCode
Open-source terminal coding assistant
Zed
High-performance, AI-native code editor
Open WebUI
Open-source, self-hosted AI chat interface
Continue
Open-source AI coding assistant for VS Code & JetBrains
See all integration guides →

Frequently asked questions

What is Grok 4.6?
Grok 4.6 is available on CrossModel as x-ai/grok-4.6. It has a 500K-token context window and can return up to 500K tokens per request.
How much does Grok 4.6 cost?
Currently $ 2.00 per 1M input tokens and $ 6.00 per 1M output tokens. Cached input is billed at $ 0.50 per 1M.
Does Grok 4.6 support tool calling and structured output?
Tool calling: Supported · Structured output (JSON): Supported · Image input: Supported · Streaming: Supported · Thinking mode: Supported
Which endpoint do I call?
For text models, OpenAI Chat Completions (/v1/chat/completions), OpenAI Responses (/v1/responses) and Anthropic Messages (/v1/messages) all work — pick whichever your SDK already speaks.
Can I try it without writing code?
Yes. Sign in and open the Playground on this page to chat with the model directly — no API key needed. Usage is billed from your wallet at the normal rate.

Chat first, integrate later

Chat first, then wire it in once you like the answers.