Claude Sonnet 4.5 vs Grok 4

Benchmarks, API pricing and specs, head to head. Data updated 2026-06-10.

Claude Sonnet 4.5

Anthropic · Sep 2025

65.5

The model that made 30-hour autonomous coding sessions real — 77.2% SWE-bench Verified and a 1M-token context beta at $3/$15.

Grok 4

xAI · Jul 2025

xAI's mid-2025 flagship, trained on the 200K-GPU Colossus cluster — a math and science reasoning standout of its generation.

The verdict

Grok 4 wins 3 of the 3 benchmarks these models share, against 0 for Claude Sonnet 4.5. Grok 4 also takes 256K of context versus 200K for Claude Sonnet 4.5.

Benchmark head-to-head 03

GPQA Diamond
83.4%87.5%
Claude Sonnet 4.5Grok 4
AIME 2025
87%91.7%
Claude Sonnet 4.5Grok 4
HLE
~17.3%25.4%
Claude Sonnet 4.5Grok 4

Specs & pricing

Claude Sonnet 4.5Grok 4
modhub Index65.5
Input price / 1M$3$3
Output price / 1M$15$15
Context window200K256K
Max output64K
Open weightsnono
Reasoning modelyesyes
Multimodal inputtext, imagetext, image
Knowledge cutoffJan 2025Jul 2025
ReleasedSep 2025Jul 2025
Example monthly cost*$52.50$52.50

* 10M input + 1.5M output tokens per month at list prices, no caching. Green = better value on that row.

Frequently asked questions

Which is better, Claude Sonnet 4.5 or Grok 4?
Grok 4 wins 3 of the 3 benchmarks these models share, against 0 for Claude Sonnet 4.5. Grok 4 also takes 256K of context versus 200K for Claude Sonnet 4.5.
Which is cheaper, Claude Sonnet 4.5 or Grok 4?
Claude Sonnet 4.5 costs $3/$15 per million input/output tokens, while Grok 4 costs $3/$15. For a typical workload of 10M input and 1.5M output tokens per month, that's $52.50 versus $52.50.
Which model is better for coding, Claude Sonnet 4.5 or Grok 4?
We don't yet track SWE-bench Verified results for both models; check their individual pages for coding-related scores.

More comparisons