Closed SourceSpaceXAIReleased 2026-08

Grok 4.6

Frontier model for long-running agents, coding, and knowledge work

Context window

500K

Input / 1M tokens

$2.00

Output / 1M tokens

$6.00

Provider

SpaceXAI

Data verified 2026-08-14

Grok 4.6 is SpaceXAI's frontier language model built on the same 1.5-trillion-parameter V9 foundation as Grok 4.5 with significantly improved supervised fine-tuning and reinforcement learning. It excels at long-running agent tasks, complex programming, and knowledge work while maintaining competitive pricing 60%+ below Claude Opus 5.

Capability index

Relative estimates (0-100) to place this model against its peers, grounded in published benchmarks.

Coding
8
Reasoning
8
Math
7
Multimodal
1
Long context
8
Speed
5
Cost efficiency
9

How to access it

Available through xAI API, Grok Build, Cursor, Grok Bot, OpenRouter, Vercel, and Cloudflare with immediate access.

Strengths

  • Frontier-level intelligence (61 on Artificial Analysis Index)
  • Exceptional long-running agent capabilities
  • Strong coding and STEM performance
  • Competitive pricing vs frontier models
  • 500K context window support

Best for developers who...

Long-running agentic tasksComplex coding and software engineeringKnowledge work and reasoningMulti-step problem solving

When to choose it (and when not to)

Reach for Grok 4.6 when...

  • When you need frontier performance at lower cost than Claude or GPT-5.6
  • For tasks requiring sustained reasoning across many steps
  • For coding-intensive workloads

Look elsewhere if...

  • If you need maximum context (older Grok 4.3 offers 1M tokens)
  • If speed is critical (slower than average relative to peers)
  • For requests consistently exceeding 200K tokens (pricing doubles)

How to use it

  • Leverage the 500K context with proper chunking for multi-document analysis
  • Use xhigh reasoning effort for complex coding or math tasks
  • Monitor 200K token threshold to avoid pricing tier jump

Quickstart

Python
from openai import OpenAI
client = OpenAI(api_key='your-key', base_url='https://api.x.ai/v1')
response = client.chat.completions.create(
  model='grok-4.6',
  messages=[{'role': 'user', 'content': 'Explain quantum computing'}]
)
print(response.choices[0].message.content)

Compatible with OpenAI SDK; configure base_url for xAI API access

API model id: grok-4.6

Benchmarks

BenchmarkScoreNotes
Artificial Analysis Intelligence Index61Matches GPT-5.6 Sol Max; trails Claude Opus 5 and Fable 5 by 1-2 points

Source: Artificial Analysis

Compare Grok 4.6

Compare Grok 4.6 with any other model

Build a comparison →
All model comparisons →

Learn the concepts