Model Comparator

Compare up to four models using the same criteria.

Gemma 4 31B (Non-reasoning)
Google
DeepSeek V3.2 Exp (Non-reasoning)
DeepSeek

Differences only

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

Model Comparator
Metrics	Gemma 4 31B (Non-reasoning) Google	DeepSeek V3.2 Exp (Non-reasoning) DeepSeek
Intelligence	13.9	13.9
Coding	33.2	—
Math	—	57.7
Speed	60 t/s	—
TTFT	1.86 s	—
Input · $ / 1M tokens	Best: $0.165	$0.28×1.7 relative to the lowest price among the compared models
Output · $ / 1M tokens	Best: $0.40	$0.42
Context window	1M	1M
Reasoning	No	No
Open weights	Yes	Yes

Intelligence

Gemma 4 31B (Non-reasoning): 13.9
DeepSeek V3.2 Exp (Non-reasoning): 13.9

Coding

Gemma 4 31B (Non-reasoning): 33.2
DeepSeek V3.2 Exp (Non-reasoning): —

Math

Gemma 4 31B (Non-reasoning): —
DeepSeek V3.2 Exp (Non-reasoning): 57.7

Speed

Gemma 4 31B (Non-reasoning): 60 t/s
DeepSeek V3.2 Exp (Non-reasoning): —

TTFT

Gemma 4 31B (Non-reasoning): 1.86 s
DeepSeek V3.2 Exp (Non-reasoning): —

Input · $ / 1M tokens

Gemma 4 31B (Non-reasoning): Best: $0.165
DeepSeek V3.2 Exp (Non-reasoning): $0.28×1.7 relative to the lowest price among the compared models

Output · $ / 1M tokens

Gemma 4 31B (Non-reasoning): Best: $0.40
DeepSeek V3.2 Exp (Non-reasoning): $0.42

Context window

Gemma 4 31B (Non-reasoning): 1M
DeepSeek V3.2 Exp (Non-reasoning): 1M

Reasoning

Gemma 4 31B (Non-reasoning): No
DeepSeek V3.2 Exp (Non-reasoning): No

Open weights

Gemma 4 31B (Non-reasoning): Yes
DeepSeek V3.2 Exp (Non-reasoning): Yes

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.

Gemma 4 31B (Non-reasoning)13.9 · $0.224
DeepSeek V3.2 Exp (Non-reasoning)13.9 · $0.315