vs
5:1
overall capability wins (0 ties across 6 scored). Magistral Small leads overall.
- Reasoning2:0
- Coding1:0
- Pricing0:3
- Speed0:2
- Specs2:1
Verdict
Magistral Small leads coding (widest gap: +31.1 on LiveCode); Ministral 3 3B offers ~2.0x larger context (256K vs 128K); Ministral 3 3B is ~7.5x cheaper on a blended token basis than Magistral Small.
- Magistral Small leads coding (widest gap: +31.1 on LiveCode)
- Ministral 3 3B offers ~2.0x larger context (256K vs 128K)
- Ministral 3 3B is ~7.5x cheaper on a blended token basis than Magistral Small
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Magistral Small
- $0.88
- Ministral 3 3B
- $0.13
Ministral 3 3B is about 7.0x cheaper on this workload.
Price per 1M tokens
Full comparison
Organization
Mistral
Mistral
Family
Mistral
Mistral
License
Apache 2.0
Apache 2.0
Open weights
Yes
Yes
Release
Jun 10, 2025
Dec 1, 2025
Knowledge cutoff
-
-
API / provider
Mistral
Mistral
Modalities
text, image → text
text, image → text
Specs
Context window
128K
256K
Max output
40K
33K
Parameters
24B
3B
Pricing
Input $/1M
$0.50
$0.10
Output $/1M
$1.50
$0.10
Blended $/1M (3∶1)
$0.75
$0.10
Speed
tok/s
90
180
TTFT (s)
0.55
0.15
Reasoning
MMLU-Pro
—
52.4
GPQA Diamond
68.2
35.8
Humanity's Last Exam
—
5.4
AIME 2025
62.8
22.0
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
—
9.0
AA-LCR v1.1
—
17.0
CritPt
—
—
MMMU-Pro
—
38.1
IFBench
—
26.8
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
—
—
SWE-bench Pro
—
—
SWE-bench Multilingual
—
—
LiveCodeBench
55.8
24.7
Terminal-Bench 2.1
—
0.0
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
—
15.3
CursorBench
—
—
SWE-Rebench
—
—
NL2Repo-Bench
—
—
DeepSWE
—
—
WebDev Arena
—
—
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
—
—
| Metric | Magistral Small | Ministral 3 3B | Delta |
|---|---|---|---|
| Identity | |||
| Organization | Mistral | Mistral | — |
| Family | Mistral | Mistral | — |
| License | Apache 2.0 | Apache 2.0 | — |
| Open weights | Yes | Yes | — |
| Release | Jun 10, 2025 | Dec 1, 2025 | — |
| Knowledge cutoff | - | - | — |
| API / provider | Mistral | Mistral | — |
| Modalities | text, image → text | text, image → text | — |
| Specs | |||
| Context window | 128K | 256K | B +128K |
| Max output | 40K | 33K | A +7K |
| Parameters | 24B | 3B | A +21 |
| Pricing | |||
| Input $/1M | $0.50 | $0.10 | B +$0.40 |
| Output $/1M | $1.50 | $0.10 | B +$1.40 |
| Blended $/1M (3∶1) | $0.75 | $0.10 | B +$0.65 |
| Speed | |||
| tok/s | 90 | 180 | B +90 |
| TTFT (s) | 0.55 | 0.15 | B +0.40 |
| Reasoning | |||
| MMLU-Pro | — | 52.4 | — |
| GPQA Diamond | 68.2 | 35.8 | A +32.4 pts |
| Humanity's Last Exam | — | 5.4 | — |
| AIME 2025 | 62.8 | 22.0 | A +40.8 pts |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | — | 9.0 | — |
| AA-LCR v1.1 | — | 17.0 | — |
| CritPt | — | — | — |
| MMMU-Pro | — | 38.1 | — |
| IFBench | — | 26.8 | — |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | — | — | — |
| SWE-bench Pro | — | — | — |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | 55.8 | 24.7 | A +31.1 pts |
| Terminal-Bench 2.1 | — | 0.0 | — |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | — | 15.3 | — |
| CursorBench | — | — | — |
| SWE-Rebench | — | — | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | — | — |
| WebDev Arena | — | — | — |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | — | — | — |
Magistral Small
- Context
- 128K / 40K out
- Parameters
- 24B
- Price
- $0.50 / $1.50
- Speed
- 90 tok/s · 0.55s TTFT
- Modalities
- text, image → text
- License
- Apache 2.0
Open-weight 24B Magistral reasoning model (Apache 2.0); deprecated on Mistral API in favor of Small 4 / Medium 3.5 reasoning modes.
Ministral 3 3B
- Context
- 256K / 33K out
- Parameters
- 3B
- Price
- $0.10 / $0.10
- Speed
- 180 tok/s · 0.15s TTFT
- Modalities
- text, image → text
- License
- Apache 2.0
Tiny Ministral 3 multimodal model - cheapest Mistral hosted tier for edge and classification.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.