Model Comparator

Compare up to four models using the same criteria.

Qwen3 30B A3B (Reasoning)
Alibaba
Devstral Small (Jul '25)
Mistral

Differences only

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

Model Comparator
Metrics	Qwen3 30B A3B (Reasoning) Alibaba	Devstral Small (Jul '25) Mistral
Intelligence	7.6	7.6
Math	Best: 72.3	29.3
Speed	113 t/s	—
TTFT	0.61 s	—
Input · $ / 1M tokens	$0.135	—
Output · $ / 1M tokens	$0.55	—
Context window	33K	Best: 256K
Reasoning	Yes	No
Open weights	Yes	Yes

Intelligence

Qwen3 30B A3B (Reasoning): 7.6
Devstral Small (Jul '25): 7.6

Math

Qwen3 30B A3B (Reasoning): Best: 72.3
Devstral Small (Jul '25): 29.3

Speed

Qwen3 30B A3B (Reasoning): 113 t/s
Devstral Small (Jul '25): —

TTFT

Qwen3 30B A3B (Reasoning): 0.61 s
Devstral Small (Jul '25): —

Input · $ / 1M tokens

Qwen3 30B A3B (Reasoning): $0.135
Devstral Small (Jul '25): —

Output · $ / 1M tokens

Qwen3 30B A3B (Reasoning): $0.55
Devstral Small (Jul '25): —

Context window

Qwen3 30B A3B (Reasoning): 33K
Devstral Small (Jul '25): Best: 256K

Reasoning

Qwen3 30B A3B (Reasoning): Yes
Devstral Small (Jul '25): No

Open weights

Qwen3 30B A3B (Reasoning): Yes
Devstral Small (Jul '25): 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.

Qwen3 30B A3B (Reasoning)7.6 · $0.239

Not plotted for lack of data on these axes: Devstral Small (Jul '25).