Is Ministral 3 3B 2512 better than Mistral Medium 3.5?

On accuracy they are statistically tied: Ministral 3 3B 2512 scored 26.2% and Mistral Medium 3.5 28.8% on the same 80 questions from open datasets. Similar accuracy; Ministral 3 3B 2512 costs less per question.

Which is cheaper, Ministral 3 3B 2512 or Mistral Medium 3.5?

Ministral 3 3B 2512 is cheaper: $0.00019 per question as billed, against $0.025. List prices: Ministral 3 3B 2512 $0.1 in / $0.1 out, Mistral Medium 3.5 $1.5 in / $7.5 out per million tokens.

Which is faster, Ministral 3 3B 2512 or Mistral Medium 3.5?

Ministral 3 3B 2512 is faster at 252 output tokens per second, against 145. Time to first token: 0.3 s for Ministral 3 3B 2512, 0.5 s for Mistral Medium 3.5.

Which has the larger context window, Ministral 3 3B 2512 or Mistral Medium 3.5?

Mistral Medium 3.5, with 262k tokens against 131k.

Head to head · tested 27 Oct 2026

Ministral 3 3B 2512 vs Mistral Medium 3.5: accuracy, speed and price

Similar accuracy; Ministral 3 3B 2512 costs less per question.

MistralMinistral 3 3B 251226.2% score · #67$0.00019 per question252 tok/s

MistralMistral Medium 3.528.8% score · #65$0.025 per question145 tok/s

Verdict

Built from the measurements below

Accuracy is too close to call: Ministral 3 3B 2512 scored 26.2% and Mistral Medium 3.5 28.8% on the same 80 questions, a gap within what chance alone could produce.
Ministral 3 3B 2512 is 128× cheaper per question as billed ($0.00019 vs $0.025), reasoning tokens included.
Ministral 3 3B 2512 generates text 1.7× faster: 252 tokens per second against 145.
Ministral 3 3B 2512 starts answering sooner: 0.3 s to the first token, against 0.5 s for Mistral Medium 3.5.
Mistral Medium 3.5 accepts more context: 262k tokens against 131k.

Which should you choose?

Choose Ministral 3 3B 2512 for the lower cost per correct answer ($0.00074 against $0.087), faster streaming, a quicker first token, shorter outputs (about 1,652 tokens per question against 3,268) and open weights you can run yourself. Choose Mistral Medium 3.5 for a 262k-token context window. The widest subject gap is in economics: Mistral Medium 3.5 answered 6 of 9 correctly, Ministral 3 3B 2512 4 of 9. With so few questions per subject, treat this as a lead to test on your own prompts, not a result.

Accuracy gap -2.5 points (z = -0.35). A winner is called only when z is at least 1.96, the usual 95% threshold. Speed, latency and price need a 10% margin.

Head to head

Same questions, same prompt, same API

Measure	Ministral 3 3B 2512	Mistral Medium 3.5	Edge
Verification scoreHigher is better	26.2%	28.8%	Too close to call
Cost per questionLower is better	$0.00019	$0.025	Ministral 3 3B 2512
Output speedHigher is better	252	145	Ministral 3 3B 2512
Time to first tokenLower is better	0.3 s	0.5 s	Ministral 3 3B 2512
End-to-end response timeLower is better	1.6 s	5.4 s	Ministral 3 3B 2512
Tokens per questionLower is better	1,652	3,268	Ministral 3 3B 2512
Context windowHigher is better	131k	262k	Mistral Medium 3.5

Side by side

Bars are scaled to the larger of the two values. The winner on each measure is marked in bold.

Every measure

Ministral 3 3B 2512 first, Mistral Medium 3.5 second

Verification scorehigher is better

Ministral 3 3B 251226.2%

Mistral Medium 3.528.8%

Cost per questionlower is better

Ministral 3 3B 2512$0.00019

Mistral Medium 3.5$0.025

Output speedhigher is better

Ministral 3 3B 2512252

Mistral Medium 3.5145

Time to first tokenlower is better

Ministral 3 3B 25120.3 s

Mistral Medium 3.50.5 s

End-to-end response timelower is better

Ministral 3 3B 25121.6 s

Mistral Medium 3.55.4 s

Tokens per questionlower is better

Ministral 3 3B 25121,652

Mistral Medium 3.53,268

Context windowhigher is better

Ministral 3 3B 2512131k

Mistral Medium 3.5262k

Score vs. cost per question

Both models in color, every other model in grey

Results by subject and dataset

Questions answered correctly. Small samples: a gap of one or two questions is noise.

Subject	Questions	Ministral 3 3B 2512	Mistral Medium 3.5
Science	30	5/30	5/30
Engineering	15	2/15	0/15
Medicine	11	4/11	3/11
Economics	9	4/9	6/9
Law	6	1/6	2/6
Humanities & social science	9	5/9	7/9
SuperGPQA dataset	50	7/50	5/50
MMLU-Pro dataset	30	14/30	18/30

Specifications and pricing

From OpenRouter's catalogue, plus the leaderboard's test dates

Specification	Ministral 3 3B 2512	Mistral Medium 3.5
Creator	Mistral	Mistral
Released	2 Dec 2025	30 Apr 2026
Weights	Open weights	—
Reasoning	No	Yes
Input	text, image	text, image, file
Output	text	text
Context window	131k tokens	262k tokens
Knowledge cutoff	—	—
List price in / out per 1M	$0.1 / $0.1	$1.5 / $7.5
Blended price per 1M (3:1)	$0.10	$3.00
Cheapest host	Mistral · $0.1 / $0.1	Mistral · $1.5 / $7.5
Hosts on OpenRouter	1	1
Last tested	26 Oct 2026	27 Oct 2026

Compare any two models →

Ministral 3 3B 2512 vs Mistral Medium 3.5: frequently asked questions

Is Ministral 3 3B 2512 better than Mistral Medium 3.5?: On accuracy they are statistically tied: Ministral 3 3B 2512 scored 26.2% and Mistral Medium 3.5 28.8% on the same 80 questions from open datasets. Similar accuracy; Ministral 3 3B 2512 costs less per question.
Which is cheaper, Ministral 3 3B 2512 or Mistral Medium 3.5?: Ministral 3 3B 2512 is cheaper: $0.00019 per question as billed, against $0.025. List prices: Ministral 3 3B 2512 $0.1 in / $0.1 out, Mistral Medium 3.5 $1.5 in / $7.5 out per million tokens.
Which is faster, Ministral 3 3B 2512 or Mistral Medium 3.5?: Ministral 3 3B 2512 is faster at 252 output tokens per second, against 145. Time to first token: 0.3 s for Ministral 3 3B 2512, 0.5 s for Mistral Medium 3.5.
Which has the larger context window, Ministral 3 3B 2512 or Mistral Medium 3.5?: Mistral Medium 3.5, with 262k tokens against 131k.