Trending

Grok 4.7 vs Claude vs GPT for business tasks

Grok 4.7 is close to the frontier on agentic office work at a much lower API price, but Claude Fable 5.1 still leads several coding and terminal benchmarks. Here's what the published numbers say, and what they don't.

What you need

• A real task to test (a report, deck, spreadsheet analysis, or code change)

Steps

1. Price: Grok 4.7 is $2 input / $6 output per 1M tokens (cache hits $0.50). In xAI's own table, GPT-5.6 Sol Max is $4/$20 and Fable 5.1 Max is $10/$50.
2. Office and knowledge work: on Artificial Analysis's AA-Briefcase, Grok 4.7 scores 1,657 Elo, just behind Claude Opus 5 and Claude Fable 5.1. xAI's table shows Fable 5.1 at 1,678 and GPT-5.6 Sol at 1,487.
3. Coding: in their native coding tools, Artificial Analysis ranks Grok 4.7 + Grok Build 4th, behind Claude Fable 5.1, GPT-6 Astra, and Claude Opus 5.
4. Long terminal work: on Terminal-Bench 4.0, xAI reports Grok 4.7 at 37.6% vs Fable 5.1 at 57.9%. Claude is clearly ahead there.
5. Legal work: xAI reports Grok 4.7 at 19.6% on the Harvey Legal Agent Benchmark vs 6.7% for Fable 5.1 and 2.5% for GPT-5.6 Sol. That's a vendor-reported number, so verify it on your own documents.
6. Token use: Artificial Analysis found Grok 4.7 (xhigh) uses about 81k output tokens per task vs 27k for GPT-6 Astra, so a lower per-token price doesn't always mean a cheaper task.
7. Context: 500k tokens, unchanged from Grok 4.6.
8. Decide by testing: run the same three real tasks through each model and score accuracy, edits needed, and cost.

Prompt to copy

Here's a real task: [paste]. Do it, then list any assumptions you made and anything you couldn't verify. I'm comparing models, so don't pad the answer.

Tips

• Several numbers come from xAI's own announcement. Prefer independent results (like Artificial Analysis) where they exist.
• Inside Grok Bot, the service selects the model. xAI says Grok 4.7 was trained to understand the Grok Bot harness.
• Benchmarks change fast. Check the dates on every source.

FAQ

Sources

• https://x.ai/news/grok-4-7
• https://artificialanalysis.ai/articles/benchmarking-grok-4-7
• https://www.reddit.com/r/singularity/comments/1wmgudm/introducing_grok_47/