Mistral Large 2 vs Mistral Medium 3

Change models
IndexAll categories, out of 100.
27.4 ±8.9 best
27.1 ±5.2
GapIndex points below the top.
Top
−0.3 · no clear gap
CoverageIndex weight that has results.
65%
95% best
SpeedOutput tokens per second.
70/s
94/s best
EloLMArena rating and rank.
1266
1370 best
ReleasedMonth of release.
2024-07
2025-05
Input / 1MUS dollars per 1M input tokens.
$2.00
$0.40 best
Output / 1MUS dollars per 1M output tokens.
$6.00
$2.00 best
1K chats2K tokens in, 500 out, each.
$7.00
$1.80 best
1K documents20K tokens in, 1K out, each.
$46.00
$10.00 best
1K agent tasks200K tokens in, 20K out, each.
$520.00
$120.00 best
ContextMaximum tokens in one request.
131K
131K
AgenticMulti-step tasks with tools.
23.2 best
18.9
CodingCode writing and repair.
—
9.3
ReasoningLogic problems and puzzles.
36.8 best
32.8
MultimodalTasks with images and text.
—
24.1
KnowledgeFacts and expert knowledge.
28.7
31.1 best
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
18.2
25.7 best
MathMath problems.
22.0
32.2 best

Benchmarks

11 shared
Only Mistral Medium 3: 3
Mistral Large 2
Mistral Medium 3
3.3%216 / 222
4.1% best199 / 222
50.2%146 / 217
59.5% best129 / 217
0.0%150 / 214
0.0%150 / 214
8.1%157 / 200
32.2% best131 / 200
48.6%178 / 199
57.8% best161 / 199
19.9% best131 / 164
18.3%139 / 164
30.7% best112 / 149
24.3%124 / 149
31.2%132 / 146
39.3% best107 / 146
47.6%50 / 88
81.6% best27 / 88
0.3%64 / 69
0.3%64 / 69
0.0%33 / 64
0.0%33 / 64