Model Comparator

Compare up to four models using the same criteria.

DeepSeek V3.2 (Non-reasoning)
DeepSeek
Qwen3.5 397B A17B (Reasoning)
Alibaba

Differences only

✓ Best published value among measured models. — Data unavailable. These metrics do not establish an overall ranking.

Model Comparator
Metrics	DeepSeek V3.2 (Non-reasoning) DeepSeek	Qwen3.5 397B A17B (Reasoning) Alibaba
Intelligence	Best: 21.5	21.4
Coding	44.2	Best: 48.2
Math	59	—
Agentic	—	8.3
Speed	—	58.9 t/s
TTFT	—	2.27 s
Input · $ / 1M tokens	Best: $0.28	$0.60×2.1 relative to the lowest price among the compared models
Output · $ / 1M tokens	Best: $0.42	$3.60×8.6 relative to the lowest price among the compared models
Context window	1M	1M
Reasoning	No	Yes
Open weights	Yes	Yes
Cost / task	—	$0.475

Intelligence

DeepSeek V3.2 (Non-reasoning): Best: 21.5
Qwen3.5 397B A17B (Reasoning): 21.4

Coding

DeepSeek V3.2 (Non-reasoning): 44.2
Qwen3.5 397B A17B (Reasoning): Best: 48.2

Math

DeepSeek V3.2 (Non-reasoning): 59
Qwen3.5 397B A17B (Reasoning): —

Agentic

DeepSeek V3.2 (Non-reasoning): —
Qwen3.5 397B A17B (Reasoning): 8.3

Speed

DeepSeek V3.2 (Non-reasoning): —
Qwen3.5 397B A17B (Reasoning): 58.9 t/s

TTFT

DeepSeek V3.2 (Non-reasoning): —
Qwen3.5 397B A17B (Reasoning): 2.27 s

Input · $ / 1M tokens

DeepSeek V3.2 (Non-reasoning): Best: $0.28
Qwen3.5 397B A17B (Reasoning): $0.60×2.1 relative to the lowest price among the compared models

Output · $ / 1M tokens

DeepSeek V3.2 (Non-reasoning): Best: $0.42
Qwen3.5 397B A17B (Reasoning): $3.60×8.6 relative to the lowest price among the compared models

Context window

DeepSeek V3.2 (Non-reasoning): 1M
Qwen3.5 397B A17B (Reasoning): 1M

Reasoning

DeepSeek V3.2 (Non-reasoning): No
Qwen3.5 397B A17B (Reasoning): Yes

Open weights

DeepSeek V3.2 (Non-reasoning): Yes
Qwen3.5 397B A17B (Reasoning): Yes

Cost / task

DeepSeek V3.2 (Non-reasoning): —
Qwen3.5 397B A17B (Reasoning): $0.475

Each reasoning level remains a separate configuration.

Score and cost

Each line joins the reasoning levels of a compared model; the large point marks the configuration in the table. The top-right corner holds the best score for the lowest cost. Select another point to compare that level.

DeepSeek V3.2 (Non-reasoning)21.5 · $0.315
Qwen3.5 397B A17B (Reasoning)21.4 · $1.35