Compare/Qwen3 Max Thinking (Preview) vs Pixtral Large

Qwen3 Max Thinking (Preview)vsPixtral Large

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

Alibaba

Qwen3 Max Thinking (Preview)

Input
$1.2/M
Output
$6/M
Speed
47 tok/s
TTFT
1.75s
Mistral

Pixtral Large

Input
$2/M
Output
$6/M
Speed
55 tok/s
TTFT
0.43s

Winner by Category

Cheaper
Qwen3 Max Thinking (Preview)
Faster (tok/s)
Pixtral Large
Lower Latency
Pixtral Large
Benchmarks (11-1)
Qwen3 Max Thinking (Preview)

Pricing Comparison

MetricQwen3 Max Thinking (Preview)Pixtral Large
Input ($/M tokens)$1.2$2
Output ($/M tokens)$6$6
Cost for 1M input + 100K output tokens:
Qwen3 Max Thinking (Preview)$1.80
Pixtral Large$2.60

Speed Comparison

Output Speed (tokens/s) — higher is better
Qwen3 Max Thinking (Preview)
47 tok/s
Pixtral Large
55 tok/s
Time to First Token (seconds) — lower is better
Qwen3 Max Thinking (Preview)
1.75s
Pixtral Large
0.43s

Benchmark Comparison

Data from Artificial Analysis API — 12 benchmarks

Intelligence Index
32.514.0
Coding Index
24.5
Math Index
82.32.3
GPQA Diamond
77.6%50.5%
MMLU-Pro
82.4%70.1%
LiveCodeBench
53.5%26.1%
AIME 2025
82.3%2.3%
MATH-500
71.4%
Humanity's Last Exam
12.0%3.6%
SciCode
38.7%29.2%
IFBench
53.8%34.5%
TerminalBench
17.4%
Qwen3 Max Thinking (Preview)11 wins
1 winsPixtral Large

Frequently Asked Questions

Which is cheaper, Qwen3 Max Thinking (Preview) or Pixtral Large?

Qwen3 Max Thinking (Preview) is cheaper overall. Its blended price (3:1 input/output ratio) is $2.40/M tokens vs $3.00/M for Pixtral Large.

Which model performs better on benchmarks?

Qwen3 Max Thinking (Preview) wins 11 out of 12 benchmarks compared to 1 for Pixtral Large. See the detailed benchmark chart above for per-category results.

Which is faster for real-time applications?

Pixtral Large generates tokens faster at 55 tok/s vs 47 tok/s. However, Pixtral Large has lower time-to-first-token (0.43s vs 1.75s).

When should I use Qwen3 Max Thinking (Preview) vs Pixtral Large?

Choose based on your priorities: Qwen3 Max Thinking (Preview) for lower cost, Qwen3 Max Thinking (Preview) for stronger benchmark performance, and Pixtral Large for faster generation. For latency-sensitive apps, check the TTFT comparison above.