Skip to main content
Radar

Keep up with the models you depend on. Follow price changes, retirements, and API updates.Follow the models you depend on.

Follow model changes

dots3-note Preview

Released Aug 14, 2026 see all recent releases

Decision reading
dots3-note Preview scores 64.4 out of 100 and ranks #40 of 232. This profile shows 31 source-displayable benchmark rows; its strongest eligible category is Agentic at #25. Published weights can be self-hosted, but infrastructure cost varies and is not a comparable API token rate.

Data as of September 10, 2026 · How the score is built

Strongest published evidence

Agentic ranks #25. Particularly useful for coding agents, browser research, and computer-use workflows.

Validate before choosing

31 published rows leave some tracked benchmark slots empty. No comparable first-party API token rate is published.

Decision snapshot

Each value carries a field reference instead of floating alone. Markers compare this model with the current ranked and priced catalog; they are not absolute quality thresholds.

Capability

64.4/100

field median 56.2

#40 of 232 ranked models

Price

Self-hosted; infrastructure cost varies

input median $0.97

No comparable first-party hosted token rate

Speed

Not measured

field median 86.5 tok/s

Time to first token not measured

Context

512Ktokens

field median 256,000

Maximum output length is tracked separately

Capability shape

Each axis shows percentile within that category’s eligible cohort. The comparison outline is the median of the six nearest public-score peers; a collapsed vertex means the category is not rank-eligible.

dots3-note Preview category percentile values

  • Agentic84th percentile
  • Coding78th percentile
  • ReasoningNot eligible
  • Knowledge77th percentile
  • MathNot eligible
  • MultilingualNot eligible
  • MultimodalNot eligible
  • Instruction following70th percentile

The dashed outline is median of 6 nearest peers.

Eligible category ranks

  1. Agentic#25/152
  2. Coding#34/151
  3. ReasoningNot ranked
  4. Knowledge#42/183
  5. MathNot ranked
  6. MultilingualNot ranked
  7. MultimodalNot ranked
  8. Inst. Following#38/123
Top decileTop quartileMid-fieldNot eligible

How much of this is verified

Coverage is split by category so a strong number never hides a thin evidence base. Verified means the row is tied to a published source; provisional rows remain visible but separate.

  1. Agentic9/9 verified
  2. Coding7/7 verified
  3. Reasoning1/1 verified
  4. Knowledge1/1 verified
  5. Math1/1 verified
  6. MultilingualNot measured
  7. Multimodal10/10 verified
  8. Inst. Following2/2 verified
Verified sourceProvisionalNot measured

Spec sheet

Each documented value carries its source. Missing fields stay visible as not sourced or not published, rather than disappearing from the page.

API model ID
Not publisheddots3-note Preview model card
Maximum output
Not sourced yet
Knowledge cutoff
Not sourced yet
Input modalities
Not sourced yet
Output modalities
Not sourced yet
Parameters
Not sourced yet
Availability
Dots Studio publishes BF16 and FP8 checkpoints for self-hosted deployment under Apache-2.0. The model card documents text, image, video, and audio input with text output; it does not publish a first-party hosted API or token rate.
Cloud regions
Not tracked yet
Lifecycle
Current
API capabilities
Tool calling, structured outputs, and batch support are not tracked yet
Prompt caching
Not documented in the pricing recorddots3-note Preview model card
Self-host
Open weights available; hardware estimate not sourced
Rate limits
Not tracked yet

Deployment options

Self-host and provider-specific paths stay separate from benchmark evidence so operating constraints are visible before a score becomes the whole decision.

Published weights are available, but BenchLM does not yet have a sourced parameter and VRAM profile for this exact model. Hardware cost estimates stay unavailable until that sizing record is complete.

Estimate VRAM from known parameters

Category score record

Scores and ranks appear only where published evidence can be displayed. The table keeps the score, weight, cohort, and evidence state together.

Category scores, ranks, weighting, benchmark coverage, and evidence status
CategoryScore
AgenticRank #25 of 152Percentile 84thWeight 22%9 benchmarksVerified59.3
CodingRank #34 of 151Percentile 78thWeight 20%7 benchmarksVerified57.2
ReasoningRank Not rankedWeight 17%1 benchmarkVerified68.3
KnowledgeRank #42 of 183Percentile 77thWeight 12%1 benchmarkVerified57.9
MathWeight 5%1 benchmarkVerifiedScore pending
MultilingualWeight 7%0 benchmarksNot measuredNot measured
MultimodalRank Not rankedWeight 12%10 benchmarksVerified66.1
Inst. FollowingRank #38 of 123Percentile 70thWeight 5%2 benchmarksVerified86.3

Benchmark ledger

Coding opens by default. The marker compares each value with the best source-verified result in the catalog; provisional leaders do not set the reference. Expand the remaining categories for every published row.

Coding7 rows
Coding benchmark values, best verified comparison, weight, and source status
SWE-bench ProScore61%Versus best verified row

Best verified: Claude Fable 5.1 · 81.2%

Gap20.2 behindWeightWeighted 25%
SWE-bench VerifiedSoftware Engineering Benchmark VerifiedScore78.4%Versus best verified row

Best verified: Claude Opus 5 · 96%

Gap17.6 behindWeightWeighted 10%
SWE MultilingualScore75.7%Versus best verified row

Best verified: Claude Opus 5 · 89.5%

Gap13.8 behindWeightWeighted 5%
CodeforcesCodeforces RatingScore3056.0Versus best verified row

Best verified: DeepSeek V4.1 Flash · 3471.0

Gap415 behindWeightDisplay only
LiveCodeBench v6Score91.5%Versus best verified row

Best verified: Sakana Fugu-Ultra · 93.2%

Gap1.7 behindWeightDisplay only
Terminal-Bench 2.1Terminal-Bench 2.1 (provider run)Score75.1%Versus best verified row

Best verified: SWE-2 · 92.8%

Gap17.7 behindWeightDisplay only
NL2RepoScore49.8%Versus best verified row

Best verified: DeepSeek V4.1 Flash · 65.4%

Gap15.6 behindWeightDisplay only
Agentic9 rows
Agentic benchmark values, best verified comparison, weight, and source status
BrowseCompScore83.3%Versus best verified row

Best verified: GPT-5.6 Sol · 92.2%

Gap8.9 behindWeightWeighted 25%
Claw-EvalScore73.4%Versus best verified row

Best verified: Ornith-1.5-397B · 81.4%

Gap8 behindWeightDisplay only
Terminal-Bench 2.1Terminal-Bench 2.1 (provider run)Score75.1%Versus best verified row

Best verified: SWE-2 · 92.8%

Gap17.7 behindWeightDisplay only
Toolathlon-VerifiedScore55.6%Versus best verified row

Best verified: Claude Opus 5 · 80.6%

Gap25 behindWeightDisplay only
skillsBenchScore52.8%Versus best verified row

Best verified: Qwen3.8 Max · 70.2%

Gap17.4 behindWeightDisplay only
APEX-AgentsScore30.8%Versus best verified row

Best verified: Grok 4.6 · 57.5%

Gap26.7 behindWeightDisplay only
HLE w/ toolsHumanity's Last Exam with toolsScore52.6%Versus best verified row

Best verified: Claude Opus 5 · 64.7%

Gap12.1 behindWeightDisplay only
DeepSearchQAScore92.1%Versus best verified row

Best verified: Claude Opus 5 · 95.0%

Gap2.9 behindWeightDisplay only
WideResearchScore78.9%Versus best verified row

Best verified: Hy4 preview · 83.9%

Gap5 behindWeightDisplay only
Reasoning1 row
Reasoning benchmark values, best verified comparison, weight, and source status
ARC-AGI-2Abstraction and Reasoning Corpus for AGI v2Score81.4%Versus best verified row

Best verified: GPT-6 Astra · 95%

Gap13.6 behindWeightWeighted 25%
Knowledge1 row
Knowledge benchmark values, best verified comparison, weight, and source status
HLEHumanity's Last ExamScore52.6%Versus best verified row

Best verified: Claude Fable 5.1 · 65%

Gap12.4 behindWeightWeighted 35%
Math1 row
Math benchmark values, best verified comparison, weight, and source status
IMOAnswerBenchScore90.9%Versus best verified row

Best verified: dots3-note Preview · 90.9%

GapBest verifiedWeightDisplay only
Multimodal10 rows
Multimodal benchmark values, best verified comparison, weight, and source status
MMMU-ProMassive Multi-discipline Multimodal Understanding ProScore79.1%Versus best verified row

Best verified: Gemini 3.5 Flash · 83.6%

Gap4.5 behindWeightWeighted 40%
SimpleVQAScore72.5%Versus best verified row

Best verified: Qwen3.7 Plus · 81.7%

Gap9.2 behindWeightDisplay only
MathVisionScore87.7%Versus best verified row

Best verified: Qwen3.8 Max · 95.2%

Gap7.5 behindWeightDisplay only
ZeroBenchScore19.0%Versus best verified row

Best verified: Muse Spark · 33.0%

Gap14 behindWeightDisplay only
CharXiv w/o toolsCharXiv Reasoning without toolsScore83.1%Versus best verified row

Best verified: Claude Mythos 5 · 88.9%

Gap5.8 behindWeightDisplay only
GDP.pdf (no tools)GDP.pdf mean criteria pass rate without toolsScore60.7%Versus best verified row

Best verified: Claude Opus 5 · 83.4%

Gap22.7 behindWeightDisplay only
PerceptionBenchPerceptionBench (Internal)Score53.4%Versus best verified row

Best verified: Qwen3.8 Max · 63.5%

Gap10.1 behindWeightDisplay only
BabyVisionScore50.0%Versus best verified row

Best verified: Qwen3.8 Max · 82.0%

Gap32 behindWeightDisplay only
MMVUMultimodal Multi-disciplinary Video UnderstandingScore79.9%Versus best verified row

Best verified: Qwen3.8 Max · 82.4%

Gap2.5 behindWeightDisplay only
VideoMMMUScore86.8%Versus best verified row

Best verified: Qwen3.8 Max · 88.7%

Gap1.9 behindWeightDisplay only
Inst. Following2 rows
Inst. Following benchmark values, best verified comparison, weight, and source status
IFBenchInstruction Following BenchmarkScore80.4%Versus best verified row

Best verified: MAI-Thinking-1 · 85%

Gap4.6 behindWeightWeighted 70%
IFEvalInstruction-Following EvalScore93.9%Versus best verified row

Best verified: Qwen3.5-27B · 95%

Gap1.1 behindWeightDisplay only

Lineage

The sequence follows explicit supersedes links. A successor's displayed score stays at least 0.1 points above its predecessor; raw benchmark rows do not move. Scores and prices remain blank when the corresponding public row or first-party rate is unavailable.

  1. Aug 14, 2026 · you are here

    dots3-note Preview

    Score 64.4 · Price not listed

preview · Preview

How to read this profile

The visual layer above carries the decisions. These notes preserve the model, ranking, coverage, and family context behind the numbers.

dots3-note Preview ranks #40 of 232 on the public leaderboard with a score of 64.41/100. It does not yet have enough sourced coverage for a verified position.

dots3-note Preview is a open weight model with a 512K context window. It uses an explicit reasoning mode, which can improve complex problem solving while adding latency and token use.

Dots Studio publishes BF16 and FP8 checkpoints for self-hosted deployment under Apache-2.0. The model card documents text, image, video, and audio input with text output; it does not publish a first-party hosted API or token rate.

Official exact-value snapshot from Dots Studio's dots3-note Preview model card. BenchLM maps only launch rows with compatible local benchmark keys, keeps Terminal-Bench 2.1 and LiveCodeBench v6 separate from older weighted lanes, and leaves ARC-AGI-3 harness variants, WorldVQA without a documented ForceAnswer setup, MME Video-v2, and other unsupported protocols outside the schema. Dots Studio reports a 280B-total, 16B-active multimodal MoE with a 512K-token context window.

31 of 435 tracked benchmark slots currently have displayable evidence. Missing categories stay blank.

Its strongest eligible category is Agentic at #25, while its lowest eligible position is Knowledge at #42. particularly useful for coding agents, browser research, and computer-use workflows.

Radar

dots3-note Preview release history

Full release history

Radar confirmed these at the source. Use dots3-note Preview in your work? Explore Radar to follow supported changes and choose your alerts.

Frequently asked questions

How does dots3-note Preview perform overall in AI benchmarks?

dots3-note Preview ranks #40 out of 232 models on the public BenchAlign leaderboard, with a score of 64.41/100. Its evidence status is Estimated, and this profile shows 31 source-displayable benchmark rows. The label describes evidence depth, not a provider quality claim; inspect category rows before choosing a workload.

Is dots3-note Preview good for knowledge and understanding?

dots3-note Preview ranks #42 out of 183 eligible models for knowledge and understanding, with a public category score of 57.9/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.

Is dots3-note Preview good for coding and programming?

dots3-note Preview ranks #34 out of 151 eligible models for coding and programming, with a public category score of 57.2/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.

Is dots3-note Preview good for mathematics?

dots3-note Preview has source-displayable benchmark coverage for mathematics, but the public category table does not assign it a rank there. The individual rows remain available for inspection. A missing category position means the evidence threshold was not met; it does not convert the model's unmeasured work into a zero.

Is dots3-note Preview good for reasoning and logic?

dots3-note Preview has source-displayable benchmark coverage for reasoning and logic, but the public category table does not assign it a rank there. The individual rows remain available for inspection. A missing category position means the evidence threshold was not met; it does not convert the model's unmeasured work into a zero.

Is dots3-note Preview good for agentic tool use and computer tasks?

dots3-note Preview ranks #25 out of 152 eligible models for agentic tool use and computer tasks, with a public category score of 59.3/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.

Is dots3-note Preview good for multimodal and grounded tasks?

dots3-note Preview has source-displayable benchmark coverage for multimodal and grounded tasks, but the public category table does not assign it a rank there. The individual rows remain available for inspection. A missing category position means the evidence threshold was not met; it does not convert the model's unmeasured work into a zero.

Is dots3-note Preview good for instruction following?

dots3-note Preview ranks #38 out of 123 eligible models for instruction following, with a public category score of 86.3/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.

Is dots3-note Preview open source?

dots3-note Preview is an open-weight model from Dots Studio. Its weights can be downloaded for local or hosted deployment, subject to the published license. Open weight does not automatically mean open source: training data and training code may remain private, and commercial restrictions can still apply.

Does dots3-note Preview have full benchmark coverage on BenchLM?

No. dots3-note Preview currently has 31 source-displayable rows across 435 tracked benchmark slots. The profile exposes published, non-generated evidence and leaves missing categories blank until an exact evaluation is available. Coverage describes how much was measured; it is not a penalty added to an individual benchmark result.

What is the context window size of dots3-note Preview?

dots3-note Preview has a documented context window of 512K. That figure is the maximum combined prompt and retained-conversation space reported for this exact model; it is not the maximum output length. The profile keeps output limits separate because providers often publish those limits independently.

Compare dots3-note Preview with every tracked model482 comparisons

Last updated September 10, 2026. Runtime fields remain blank until a sourced snapshot exists.

Watch dots3-note Preview in the weekly brief

Get one weekly email when material rank, price, availability, or benchmark evidence changes are worth revisiting.

Read a sample issue

Join 2,000+ readers.