Qwen 3.6 Max (preview) vs Qwen3.7 Plus

Change models
IndexAll categories, out of 100.
56.4 ±5.0
56.4 ±2.3
GapIndex points below the top.
Top
Top
CoverageIndex weight that has results.
85%
100% best
SpeedOutput tokens per second.
75/s best
50/s
EloLMArena rating and rank.
1446
1454 best
ReleasedMonth of release.
2026-04
2026-06
Input / 1MUS dollars per 1M input tokens.
$1.03
$0.32 best
Output / 1MUS dollars per 1M output tokens.
$6.16
$1.28 best
Batch output / 1MBatch price per 1M output tokens.
—
$0.92
1K chats2K tokens in, 500 out, each.
$5.14
$1.28 best
1K documents20K tokens in, 1K out, each.
$26.70
$7.68 best
1K agent tasks200K tokens in, 20K out, each.
$328.64
$89.60 best
ContextMaximum tokens in one request.
262K
1M best
AgenticMulti-step tasks with tools.
58.5 best
56.3
CodingCode writing and repair.
56.3
57.5 best
ReasoningLogic problems and puzzles.
52.5
56.8 best
MultimodalTasks with images and text.
—
61.6
KnowledgeFacts and expert knowledge.
58.7 best
54.2
MultilingualTasks in many languages.
—
53.3
InstructionTasks with strict prompt rules.
59.6 best
57.2
MathMath problems.
52.7 best
51.2

Benchmarks

16 shared
Only Qwen 3.6 Max (preview): 12Only Qwen3.7 Plus: 60
Qwen 3.6 Max (preview)
Qwen3.7 Plus
87.4% best67 / 220
81.8%85 / 220
30.8%72 / 216
35.6% best61 / 216
80.7% best37 / 211
73.0%98 / 211
3.7%91 / 207
9.1% best69 / 207
91.1% best61 / 203
80.0%89 / 203
88.8%48 / 193
90.0% best36 / 193
95.9% best13 / 143
93.0%31 / 143
76.6%10 / 140
78.0% best7 / 140
19.0% best49 / 83
17.0%54 / 83
57.3%43 / 83
57.6% best42 / 83
65.4%12 / 40
70.3% best4 / 40
NL2RepoCoding
42.9% best22 / 36
41.1%26 / 36
47.0%13 / 35
51.3% best7 / 35
43.9%17 / 22
47.0% best11 / 22
73.9% best3 / 21
71.4%6 / 21
59.0%3 / 10
61.8% best2 / 10