Compare/Claude 3.7 Sonnet (Non-reasoning) vs Grok 3

Claude 3.7 Sonnet (Non-reasoning)vsGrok 3

Side-by-side comparison of pricing, 12 benchmarks, and generation speed.

Anthropic

Claude 3.7 Sonnet (Non-reasoning)

Input
$3/M
Output
$15/M
Speed
TTFT
xAI

Grok 3

Input
$3/M
Output
$15/M
Speed
70 tok/s
TTFT
0.33s

Winner by Category

Cheaper
Tie
Faster (tok/s)
Grok 3
Lower Latency
Claude 3.7 Sonnet (Non-reasoning)
Benchmarks (5-7)
Grok 3

Pricing Comparison

MetricClaude 3.7 Sonnet (Non-reasoning)Grok 3
Input ($/M tokens)$3$3
Output ($/M tokens)$15$15
Cost for 1M input + 100K output tokens:
Claude 3.7 Sonnet (Non-reasoning)$4.50
Grok 3$4.50

Speed Comparison

Output Speed (tokens/s) — higher is better
Claude 3.7 Sonnet (Non-reasoning)
Grok 3
70 tok/s
Time to First Token (seconds) — lower is better
Claude 3.7 Sonnet (Non-reasoning)
Grok 3
0.33s

Benchmark Comparison

Data from Artificial Analysis API — 12 benchmarks

Intelligence Index
30.825.2
Coding Index
26.719.8
Math Index
21.058.0
GPQA Diamond
65.6%69.3%
MMLU-Pro
80.3%79.9%
LiveCodeBench
39.4%42.5%
AIME 2025
21.0%58.0%
MATH-500
85.0%87.0%
Humanity's Last Exam
4.8%5.1%
SciCode
37.6%36.8%
IFBench
44.0%46.9%
TerminalBench
21.2%11.4%
Claude 3.7 Sonnet (Non-reasoning)5 wins
7 winsGrok 3

Frequently Asked Questions

Which is cheaper, Claude 3.7 Sonnet (Non-reasoning) or Grok 3?

Both models have similar pricing. Check the detailed breakdown above for input vs output token costs.

Which model performs better on benchmarks?

Grok 3 wins 7 out of 12 benchmarks compared to 5 for Claude 3.7 Sonnet (Non-reasoning). See the detailed benchmark chart above for per-category results.

Which is faster for real-time applications?

Grok 3 generates tokens faster at 70 tok/s vs 0 tok/s. Claude 3.7 Sonnet (Non-reasoning) also has lower time-to-first-token (0.00s vs 0.33s).

When should I use Claude 3.7 Sonnet (Non-reasoning) vs Grok 3?

Choose based on your priorities: both are similarly priced, Grok 3 for stronger benchmark performance, and Grok 3 for faster generation. For latency-sensitive apps, check the TTFT comparison above.