DeepSeek V3.2 vs Qwen3 Max

Change models
Qwen3 MaxAlibaba
IndexAll categories, out of 100.
44.3 ±4.9
44.5 ±5.2 best
GapIndex points below the top.
−0.2 · no clear gap
Top
CoverageIndex weight that has results.
85%
85%
SpeedOutput tokens per second.
36/s
62/s best
EloLMArena rating and rank.
1425 best
1413
ReleasedMonth of release.
2025-12
2025-09
Input / 1MUS dollars per 1M input tokens.
$0.26 best
$0.78
Output / 1MUS dollars per 1M output tokens.
$0.38 best
$3.90
1K chats2K tokens in, 500 out, each.
$0.71 best
$3.51
1K documents20K tokens in, 1K out, each.
$5.58 best
$19.50
1K agent tasks200K tokens in, 20K out, each.
$59.60 best
$234.00
ContextMaximum tokens in one request.
164K
262K best
AgenticMulti-step tasks with tools.
44.2
45.9 best
CodingCode writing and repair.
48.3 best
45.9
ReasoningLogic problems and puzzles.
42.5 best
40.1
MultimodalTasks with images and text.
—
—
KnowledgeFacts and expert knowledge.
43.4
48.0 best
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
34.9 best
30.5
MathMath problems.
46.4 best
45.8

Benchmarks

24 shared
Only DeepSeek V3.2: 14Only Qwen3 Max: 9
DeepSeek V3.2
Qwen3 Max
77.3% best99 / 220
72.6%113 / 220
11.2%132 / 216
11.9% best130 / 216
45.7%157 / 211
50.0% best148 / 211
0.9% best119 / 207
0.0%145 / 207
68.4%106 / 203
73.3% best97 / 203
75.1%112 / 193
76.4% best106 / 193
24.0%111 / 170
24.4% best108 / 170
66.0%135 / 149
81.9% best78 / 149
78.9% best68 / 143
74.3%73 / 143
69.9%91 / 142
78.2% best76 / 142
49.0% best76 / 140
44.1%85 / 140
76.3%78 / 135
79.5% best69 / 135
MMLU ProKnowledge
83.1%74 / 135
84.4% best62 / 135
CorpFin v2Finance
47.9%110 / 131
55.9% best91 / 131
AIMEMath
64.8%55 / 95
81.0% best46 / 95
MedQAHealthcare
89.5% best51 / 95
87.4%57 / 95
MGSMMultilingual
90.9%31 / 74
92.1% best21 / 74
34.8% best42 / 66
24.7%52 / 66
IOI v1Coding
14.4%31 / 61
15.7% best28 / 61
0.0%51 / 51
44.3% best39 / 51