Compare/Qwen3.5 27B (Non-reasoning) vs Grok 4.3 (high)

Qwen3.5 27B (Non-reasoning)vsGrok 4.3 (high)

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

Alibaba

Qwen3.5 27B (Non-reasoning)

Input
$0.28/M
Output
$2.5/M
Speed
89 tok/s
TTFT
1.40s
xAI

Grok 4.3 (high)

Input
$1.25/M
Output
$2.5/M
Speed
85 tok/s
TTFT
9.52s

Winner by Category

Cheaper
Qwen3.5 27B (Non-reasoning)
Faster (tok/s)
Qwen3.5 27B (Non-reasoning)
Lower Latency
Qwen3.5 27B (Non-reasoning)
Benchmarks (0-7)
Grok 4.3 (high)

Pricing Comparison

MetricQwen3.5 27B (Non-reasoning)Grok 4.3 (high)
Input ($/M tokens)$0.28$1.25
Output ($/M tokens)$2.5$2.5
Cost for 1M input + 100K output tokens:
Qwen3.5 27B (Non-reasoning)$0.53
Grok 4.3 (high)$1.50

Speed Comparison

Output Speed (tokens/s) — higher is better
Qwen3.5 27B (Non-reasoning)
89 tok/s
Grok 4.3 (high)
85 tok/s
Time to First Token (seconds) — lower is better
Qwen3.5 27B (Non-reasoning)
1.40s
Grok 4.3 (high)
9.52s

Benchmark Comparison

Data from Artificial Analysis API — 12 benchmarks

Intelligence Index
37.253.2
Coding Index
33.441.0
Math Index
GPQA Diamond
84.2%90.1%
MMLU-Pro
LiveCodeBench
AIME 2025
MATH-500
Humanity's Last Exam
13.2%35.0%
SciCode
36.7%47.3%
IFBench
46.9%81.3%
TerminalBench
31.8%37.9%
Qwen3.5 27B (Non-reasoning)0 wins
7 winsGrok 4.3 (high)

Frequently Asked Questions

Which is cheaper, Qwen3.5 27B (Non-reasoning) or Grok 4.3 (high)?

Qwen3.5 27B (Non-reasoning) is cheaper overall. Its blended price (3:1 input/output ratio) is $0.83/M tokens vs $1.56/M for Grok 4.3 (high).

Which model performs better on benchmarks?

Grok 4.3 (high) wins 7 out of 12 benchmarks compared to 0 for Qwen3.5 27B (Non-reasoning). See the detailed benchmark chart above for per-category results.

Which is faster for real-time applications?

Qwen3.5 27B (Non-reasoning) generates tokens faster at 89 tok/s vs 85 tok/s. Qwen3.5 27B (Non-reasoning) also has lower time-to-first-token (1.40s vs 9.52s).

When should I use Qwen3.5 27B (Non-reasoning) vs Grok 4.3 (high)?

Choose based on your priorities: Qwen3.5 27B (Non-reasoning) for lower cost, Grok 4.3 (high) for stronger benchmark performance, and Qwen3.5 27B (Non-reasoning) for faster generation. For latency-sensitive apps, check the TTFT comparison above.