Inference
EN
Dots Studio

dots3-note Preview

dots3-note Preview is a reasoning model from Dots Studio in the dots3-note family. 23 benchmarks count toward its index, in 7 categories.

  • Open weights
  • Reasoning
  • Input: text
  • Released 2026-08
RankedEnough results for an index.
IndexAll categories, out of 100.63.4±3.9Median 51.0
CoverageIndex weight that has results.95%
SpeedOutput tokens per second.—Median 67/s
First tokenSeconds to the first output token.—Median 1.3 s
Input / 1MUS dollars per 1M input tokens.FreeMedian $0.60
Output / 1MUS dollars per 1M output tokens.FreeMedian $2.55
ContextMaximum tokens in one request.512KMedian 262K
EloLMArena rating and rank.—Median 1422

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.8 counted63.356 / 218
CodingCode writing and repair.4 counted60.369 / 209
ReasoningLogic problems and puzzles.1 counted74.5—
MultimodalTasks with images and text.6 counted59.244 / 84
KnowledgeFacts and expert knowledge.1 counted77.2—
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.2 counted58.217 / 63
MathMath problems.1 counted62.3—

Results

23 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
IMOAnswerBenchMath90.9%1 / 10
LiveCodeBench v6Coding91.5%7 / 59
Instruction-Following EvalInstruction93.9%10 / 70
Claw-EvalAgentic73.4%6 / 42
ARC-AGI-2 (semi-private)Reasoning76.8%22 / 131
Instruction Following BenchmarkInstruction80.4%11 / 57
Humanity's Last ExamKnowledge52.6%18 / 77
Software Engineering Benchmark VerifiedCoding78.4%27 / 114
DeepSearchQAAgentic92.1%8 / 24
SimpleVQAMultimodal72.5%6 / 18
NL2RepoCoding49.8%12 / 36
Massive Multi-discipline Multimodal Understanding ProMultimodal79.1%18 / 53
SWE-bench ProCoding61.0%31 / 83
BrowseCompAgentic83.3%24 / 63
WideResearchAgentic78.9%7 / 16
MathVisionMultimodal87.7%12 / 24
Terminal-Bench 2.1 · BenchLMAgentic75.1%44 / 86
Humanity's Last Exam with toolsAgentic52.6%23 / 36
CharXiv Reasoning without toolsMultimodal83.1%11 / 17
APEX-AgentsAgentic30.8%9 / 13
ZeroBenchMultimodal19.0%8 / 10
BabyVisionMultimodal50.0%9 / 11
Toolathlon-VerifiedAgentic55.6%19 / 23
VideoMMMUMultimodal86.8%—
Multimodal Multi-disciplinary Video UnderstandingMultimodal79.9%—
GDP.pdf mean criteria pass rate without toolsMultimodal60.7%—
PerceptionBench (Internal)Multimodal53.4%—

Sources

BenchLMCC BY-NC 4.0ARC Prize FoundationTerms restrict reuse

History

5 releases
Index64.5 → 63.4
23 Oct 202612 Nov 2026
Output / 1MFree → Free
23 Oct 202612 Nov 2026

Questions

How much does dots3-note Preview cost?
Input tokens are free, output tokens are free.
How good is dots3-note Preview?
Its index is 63.4 out of 100. That is place 57 of 283 ranked models.
What is dots3-note Preview best at?
Agentic: 63.3 out of 100, place 56 of 218.
What is the context window of dots3-note Preview?
512K tokens.
Is dots3-note Preview open source?
Yes. Its weights are published.
When was dots3-note Preview released?
August 2026.

Badge

dots3-note Preview on Inference360: place 57 of 283, index 63.4
HTML
<a href="https://inference360.ai/models/dots3-note-preview"><img src="https://inference360.ai/badges/dots3-note-preview.svg" alt="dots3-note Preview on Inference360: place 57 of 283, index 63.4" height="20"></a>
Markdown
[![dots3-note Preview on Inference360: place 57 of 283, index 63.4](https://inference360.ai/badges/dots3-note-preview.svg)](https://inference360.ai/models/dots3-note-preview)

Compare with

Claude Opus 4.7 (Adaptive)63.2GLM-5.3-Flash63.1Grok 4.563.7Qwen3.8-Omni-Flash62.7Gemini 3.5 Flash64.2