AI BENCHY
AD
Track all your projects in one dashboard. Get 📊stats, 🔥heatmaps and 👀recordings in one self-hosted dashboard.
uxwizz.com
#188

Dots 3 Note Preview

Dots Studio Release: 2026-08-14 Tested on: 2026-08-15 21:56 dots-studio/dots-3-note-preview::medium
280B total (16B active)MoEOpen source
(high) (medium) (low) (none)

Summary

Dots 3 Note Preview scores 5.9 on AI BENCHY and ranks #188. It has 10.0 reliability, a 57.6% pass rate, $0.000 total cost, and 67.13s average response time.

What makes Dots 3 Note Preview unique: It stands out most in Combined, where it ranks #1, while General Intelligence is its weakest area at #13. Its total benchmark cost is unusually low for its score range. It uses unusually many reasoning tokens, which can help explain its slower or more expensive runs.

Model facts

Researched on 2026-08-15

Reported
Parameters
280B total (16B active)
Architecture
MoE
Availability
Open source
License
Apache-2.0

The official model card publishes both weights and implementation code.

Score

5.9

Consistency

7.1

Total Output Tokens

679,539

Total Input Tokens

94,639

Input Price

$0.000 / 1M

Output Price

$0.000 / 1M

Tests Correct

Wrong Tests: 14

Attempt pass rate: 57.6%

Flaky tests

8

Flaky tests had mixed outcomes across runs (at least one pass and one fail).

Response Time (avg)

67.13s

Response Time (max): 459.19s

Response Time (total): 1476.95s

Hamster playing table tennis

Prompt: Create a detailed SVG illustration of a hamster playing table tennis.

#188 Dots 3 Note Preview

medium
Cost
$0.000
Time
585.4s
Tokens
12,198 tok

Charts

Choose the first model, then click a second model to open a side-by-side page.

Total Output Tokens

Score vs Total Output Tokens

Quick Compare

Category Breakdown

Category Score Consistency Tests Correct
Anti-AI Tricks 8.7 7.9
Coding 4.2 5.1
Combined 9.9 10.0
Data parsing and extraction 4.5 1.4
Domain specific 3.6 7.2
General Intelligence 3.8 9.5
Instructions following 8.5 6.8
Puzzle Solving 5.3 7.2
Tool Calling 10.0 10.0
Trivia 3.0 10.0

Compared models