Aino — Selection Console V3.2.5.11

Aino Home

Selection Console · V3.2.5.11

DeepSeek V4.1 Flash

Paid, metered API · model: deepseek-flash

Input price	$0.15 / 1M tokens (off-peak)
Output price	$0.60 / 1M tokens (off-peak)
Peak-hour pricing	2× both
Context window	1,000,000 tokens
Max output	384,000 tokens
Reasoning	Extended "High" effort — on by default

Peak vs. Off-Peak Details

Peak Schedule: 7 total hours each weekday (Monday–Friday).
Off-Peak Schedule: All remaining weekday hours, plus the entire weekend (Saturday and Sunday) and Chinese public holidays.

Local Conversion (EET / Helsinki Time, UTC+3):
Peak Window 1: 04:00 – 07:00
Peak Window 2: 09:00 – 13:00

With reasoning enabled, the model spends part of its output budget "thinking" before answering — this console floors the token ceiling at 8,000 per request to avoid empty responses.

Typical assessment cost: well under a cent, even with images attached
No free tier — this is billed against your prepaid DeepSeek balance

This is a paid account, not a shared rate-limited free tier, so day-to-day use shouldn't hit hard caps the way earlier free-tier providers did. Track actual spend and any account-specific limits at platform.deepseek.com.

Glossary of signals

What each of the 12 signals is actually measuring.

Animal safety	Physical or chemical hazard risk to the pet
Human safety	Hazard risk to the humans handling or near it
Health & wellbeing	Support for the pet's physical health (posture, warmth, pressure relief)
Comfort	How pleasant/soft the product feels to the pet
Cleanliness	How well it resists dirt, odor, or staining over time
Easy maintenance	How easy it is to wash, clean, or upkeep
Convenience	Ease of use, storage, and handling for the owner
Interaction	How it facilitates engagement between pet and owner
Agency	How much freedom/control the pet has (movement, choice)
Organisation	How well it helps keep the home tidy
Space efficiency	How compact or space-saving the product is
Aesthetics	Visual style and appeal

The Gate & DNA panel groups some of these into composites: Pet satisfaction (health & wellbeing + comfort + a share of interaction), Human satisfaction (convenience + agency + a share of interaction), and Orderliness (organisation + space efficiency). Animal safety, human safety, cleanliness, and aesthetics stand alone.

Glossary of assessment results

What each final Selection Engine result means. The Gate is a hard constraint and can override the Selection Score band.

SELECT	Gate passed and Selection Score is ≥80. The product is ready to be selected under the current evidence and configuration.
TEST	Gate passed and Selection Score is ≥65 but below 80, or a raw SELECT result is capped at TEST because animal/human safety evidence is incomplete or below its floor.
HOLD	The product is not ready for a selection decision because the Gate requires more evidence, such as insufficient coverage, low confidence, or partial DNA. A score band of 50–64.9 also maps to HOLD when the Gate passes.
REJECT	A hard Gate condition failed, such as a critical safety hazard or failed DNA threshold, or the Selection Score is below 50 when the Gate passes.

Important: Product assessment and transaction economics are separate decisions. A pricing result such as REVIEW or NOT SUPPORTED does not by itself change the product assessment result.

Glossary of Gate & DNA dimensions

What the Gate checks, what each DNA dimension measures, and the configured floors used by the Selection Engine.

Gate

Gate status	PASS, HOLD, or REJECT. The Gate is a hard constraint and overrides the Selection Score band.
Hard safety trigger	REJECT if animal or human safety evidence is strongly negative (polarity < 0 and strength ≥ 3), or if verified injury, manufacturer hazard warning, or structural failure hazard is present.
Evidence sufficiency	HOLD when signal coverage is below the limited-evidence threshold or aggregate confidence is below 40%.
DNA thresholds	Non-safety DNA dimensions are checked against their configured floors. A missing dimension places the Gate on HOLD; a dimension below its floor fails the Gate.
Safety verification	Human and animal safety dimensions are tracked as safety unverified when evidence is present but the dimension remains below its floor. This does not itself fail the Gate, but it can cap a SELECT decision at TEST.
Overall DNA	HOLD when Overall DNA is missing or partial; REJECT only when complete Overall DNA is below the configured overall floor.

DNA dimensions

Human safety	Directly reflects the Human safety signal. Floor: 80.
Animal safety	Directly reflects the Animal safety signal. Floor: 85.
Cleanliness	Directly reflects the Cleanliness signal. Floor: 60.
Aesthetics	Directly reflects the Aesthetics signal. Floor: 50.
Pet satisfaction	Composite of Health & wellbeing (34%), Comfort (33%), and an Interaction share (33%). Interaction contributes half of its normalized score to this dimension. Floor: 50.
Human satisfaction	Composite of Convenience (34%), Agency (33%), and an Interaction share (33%). Interaction contributes half of its normalized score to this dimension. Floor: 50.
Orderliness	Composite of Organisation (50%) and Space efficiency (50%). Floor: 55.
Overall DNA (A_D)	Weighted composite of the DNA dimensions, using the configured weights and renormalising over dimensions that have values. Floor: 70.

Configured DNA weights

Animal safety	25%
Human safety	20%
Pet satisfaction	15%
Human satisfaction	15%
Cleanliness	10%
Aesthetics	7.5%
Orderliness	7.5%