Skip to content
AIMODELRACE
xAILanguage modelNew

Grok 4.7

xAI's best coder yet: elite math, live X knowledge, aggressive pricing.

Overall Intelligence

#11

of 56
4.3 pts behind #1

0

races led

0

weeks at #1

84.5

score

Released
Sep 21, 2026
4 wk ago
Context
1M
128K max output
Price
$2 / $6
per 1M in/out · $3.00 blended
Speed
112 tok/s
4s to first token
Modalities
text · image
outputs text
Weights
Closed
API and apps

The 30-second version

xAI's best coder yet, with elite math, live knowledge from X and aggressive pricing. Grok 4.7 is a value pick for technical work and real-time information.

Key takeaways
  • 1Top-three on our math races at $2 / $6 per million tokens.
  • 2Real-time awareness of news and trends via X.
  • 3Big coding improvement over Grok 4.6.
  • 4Less polished writing and more variable on long agent runs than the leaders.

What you need to know

Purpose

A fast-moving frontier model for technical users who want strong reasoning and up-to-the-minute information at a low price.

What xAI optimised for

Raw reasoning and coding capability scaled on xAI's large training clusters, with live data integration as a differentiator.

Unique strengths

  • Live X data and trend awareness.
  • Strong competition math.
  • Low output price.
Insight
How it differs from rivals. Cheaper output than any other frontier contender. Versus GPT-6.1 Sol: better at math, weaker at writing and instruction following.

Where it ranks

Current snapshot · illustrative scores · click any race to watch it

CodingHard ReasoningMathScienceWriting & LanguageMultimodalTool Use & AgentsLong ContextTrust & FactualitySpeed & Cost

Grok 4.7Best on each axis

Overall Intelligence over time

#1 frontierGrok 4.7Muse WonderSonnet 5.5
55.060.065.070.075.080.085.090.02025Apr ’25Jul ’25Oct ’252026Apr ’26Jul ’26Oct ’26

Best use cases

01

Math and quantitative work

Problem solving, modelling and derivations.

02

Real-time research

What are people saying about this right now?

03

Budget coding

Solid coding help at a lower price than the leaders.

Strengths

  • Math

    Near the top of the competition-math race.

  • Price

    $6 per million output tokens is very aggressive for this tier.

Weaknesses

  • Consistency

    More variance between runs on long, agentic tasks.

  • Writing polish

    Behind GPT-6.1 Sol and Claude on nuanced prose.

When to choose it, and when not to

Pick something else when…

  • Brand-sensitive writing.
  • Mission-critical long autonomous runs.

Update history

The distilled version first; open any release for the deeper breakdown.