dots3-note Preview vs Qwen3.5-27B

Change models
dots3-note PreviewDots Studio
IndexAll categories, out of 100.
63.4 ±3.9 best
51.3 ±4.3
GapIndex points below the top.
Top
−12.1
CoverageIndex weight that has results.
95% best
90%
SpeedOutput tokens per second.
—
69/s
EloLMArena rating and rank.
—
1409
ReleasedMonth of release.
2026-08
2026-03
Input / 1MUS dollars per 1M input tokens.
Free best
$0.26
Output / 1MUS dollars per 1M output tokens.
Free best
$2.60
1K chats2K tokens in, 500 out, each.
Free best
$1.82
1K documents20K tokens in, 1K out, each.
Free best
$7.80
1K agent tasks200K tokens in, 20K out, each.
Free best
$104.00
ContextMaximum tokens in one request.
512K best
262K
AgenticMulti-step tasks with tools.
63.3 best
50.3
CodingCode writing and repair.
60.3 best
52.7
ReasoningLogic problems and puzzles.
74.5 best
50.5
MultimodalTasks with images and text.
59.2 best
53.9
KnowledgeFacts and expert knowledge.
77.2 best
49.0
MultilingualTasks in many languages.
—
51.6
InstructionTasks with strict prompt rules.
58.2 best
57.1
MathMath problems.
62.3
—

Benchmarks

7 shared
Only dots3-note Preview: 20Only Qwen3.5-27B: 22
dots3-note Preview
Qwen3.5-27B
78.4% best27 / 114
72.4%60 / 114
93.9%10 / 70
95.0% best4 / 70
BrowseCompAgentic
83.3% best24 / 63
61.0%51 / 63
91.5% best7 / 59
80.7%27 / 59
79.1% best18 / 53
75.0%38 / 53
MathVisionMultimodal
87.7% best12 / 24
86.0%16 / 24
79.9%
73.3%