vs
9:1
overall capability wins (0 ties across 10 scored). Devstral 2 leads overall.
- Reasoning4:1
- Coding2:0
- Pricing0:3
- Speed0:2
- Specs3:0
Verdict
Devstral 2 leads coding (widest gap: +27.4 on SWE-Rebench); Devstral 2 offers ~2.0x larger context (256K vs 128K); Devstral 2 allows ~7.8x more max output (64K vs 8K).
- Devstral 2 leads coding (widest gap: +27.4 on SWE-Rebench)
- Devstral 2 offers ~2.0x larger context (256K vs 128K)
- Devstral 2 allows ~7.8x more max output (64K vs 8K)
- Devstral Small is ~6.5x cheaper on a blended token basis than Devstral 2
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Devstral 2
- $0.90
- Devstral Small
- $0.14
Devstral Small is about 6.4x 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
Dec 1, 2025
May 21, 2025
Knowledge cutoff
-
-
API / provider
Mistral
Mistral
Modalities
text → text
text → text
Specs
Context window
256K
128K
Max output
64K
8K
Parameters
123B
24B
Pricing
Input $/1M
$0.40
$0.07
Output $/1M
$2
$0.28
Blended $/1M (3∶1)
$0.80
$0.12
Speed
tok/s
85
100
TTFT (s)
0.35
0.28
Reasoning
MMLU-Pro
76.2
—
GPQA Diamond
59.4
41.4
Humanity's Last Exam
3.6
3.8
AIME 2025
36.7
—
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
20.8
14.7
AA-LCR v1.1
32.3
18.7
CritPt
—
—
MMMU-Pro
—
—
IFBench
38.1
34.6
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
—
46.8
SWE-bench Pro
—
—
SWE-bench Multilingual
—
—
LiveCodeBench
44.8
—
Terminal-Bench 2.1
30.3
—
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
32.8
24.3
CursorBench
—
—
SWE-Rebench
42.0
14.6
NL2Repo-Bench
—
—
DeepSWE
—
—
WebDev Arena
1,194
—
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
—
—
| Metric | Devstral 2 | Devstral Small | Delta |
|---|---|---|---|
| Identity | |||
| Organization | Mistral | Mistral | — |
| Family | Mistral | Mistral | — |
| License | Apache 2.0 | Apache 2.0 | — |
| Open weights | Yes | Yes | — |
| Release | Dec 1, 2025 | May 21, 2025 | — |
| Knowledge cutoff | - | - | — |
| API / provider | Mistral | Mistral | — |
| Modalities | text → text | text → text | — |
| Specs | |||
| Context window | 256K | 128K | A +128K |
| Max output | 64K | 8K | A +56K |
| Parameters | 123B | 24B | A +99 |
| Pricing | |||
| Input $/1M | $0.40 | $0.07 | B +$0.33 |
| Output $/1M | $2 | $0.28 | B +$1.72 |
| Blended $/1M (3∶1) | $0.80 | $0.12 | B +$0.68 |
| Speed | |||
| tok/s | 85 | 100 | B +15 |
| TTFT (s) | 0.35 | 0.28 | B +0.07 |
| Reasoning | |||
| MMLU-Pro | 76.2 | — | — |
| GPQA Diamond | 59.4 | 41.4 | A +18.0 pts |
| Humanity's Last Exam | 3.6 | 3.8 | B +0.2 pts |
| AIME 2025 | 36.7 | — | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | 20.8 | 14.7 | A +6.1 pts |
| AA-LCR v1.1 | 32.3 | 18.7 | A +13.6 pts |
| CritPt | — | — | — |
| MMMU-Pro | — | — | — |
| IFBench | 38.1 | 34.6 | A +3.5 pts |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | — | 46.8 | — |
| SWE-bench Pro | — | — | — |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | 44.8 | — | — |
| Terminal-Bench 2.1 | 30.3 | — | — |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | 32.8 | 24.3 | A +8.5 pts |
| CursorBench | — | — | — |
| SWE-Rebench | 42.0 | 14.6 | A +27.4 pts |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | — | — |
| WebDev Arena | 1,194 | — | — |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | — | — | — |
Devstral 2
- Context
- 256K / 64K out
- Parameters
- 123B
- Price
- $0.40 / $2
- Speed
- 85 tok/s · 0.35s TTFT
- Modalities
- text → text
- License
- Apache 2.0
Open-weight agentic coding model for autonomous software engineering (successor path: Medium 3.5).
Devstral Small
- Context
- 128K / 8K out
- Parameters
- 24B
- Price
- $0.07 / $0.28
- Speed
- 100 tok/s · 0.28s TTFT
- Modalities
- text → text
- License
- Apache 2.0
Open coding agent model from Mistral — lighter sibling to Devstral 2 for IDE workflows.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.