Kingy AI
Subscribe, 2.1M subscribersWork with Kingy

AI Model Profile

Kimi K3

Kimi K3 is a multimodal reasoning model with text, image and video input and a 1,048,576-token context window.

Before you buy

Best for: Creators building coding, research or structured-content workflows who value open weights and can accept slower reasoning.

Price: US/USD Kimi API: US$3/M input; US$15/M output; cache writes extra

Included: Pay-as-you-go API; minimum US$1 top-up unlocks K3. Cached input is US$0.30/M tokens; cache writes cost US$3/M for 5 minutes or US$6/M for 1 hour. Taxes extra.

Main limits: Thinking is always on. Rate limits depend on account tier. Self-hosting this 2.8-trillion-parameter model needs substantial infrastructure; available weights do not mean a laptop-sized download or a permissive standard license.

What Kingy showed: Kingy’s September 2 hosted study completed 24/24 Kimi K3 requests: mean score 94.6, 87.5% meeting the 80-point threshold, and 15.3-second median latency.

Buying details checked 2026-10-22 with region and currency stated above. Historical demonstrations and the broader fact ledger keep their own dates.

Check official plans · See the demonstration and disclosure · Compare these workflows

Official model page Official docs Compare models
Family
Kimi K3
Release date
Unknown
Status
Kimi K3 API, official model guide and downloadable weights checked October 22, 2026. International API access requires a minimum US$1 top-up.
Context window
1,048,576 tokens
Output limit
Unknown
API
yes
Open weights
yes
Local/self-hosted
yes
Pricing
USD per million tokens: input $3; output $15; cached input $0.30. Cache writes: 5-minute TTL $3; 1-hour TTL $6. Taxes extra.
Evidence state
Source-verified

Benchmark Caveat

Kingy September 2 hosted OpenRouter study: 24 requests, 12 tasks twice, medium effort, 4096-token cap, five providers. Small dated workload; not current native-API or self-hosted performance.

Best for

Creators building coding, research or structured-content workflows who value open weights and can accept slower reasoning.

Kingy AI Product Facts

Kimi K3

Current statusUnknown

Company
Not yet reviewed
Primary job
Not yet reviewed
Audience
Not yet reviewed
Evidence
Source-backed; coverage expanding
Latest source check
October 23, 2026
See all tracked product factsPricing, platforms, dependencies, data claims, regions, and timeline

Sources and checks

Coverage
1 of 88 tracked fields currently verified
Sources
1

Source-check dates apply to the tracked facts. Editorial updates elsewhere on this page may have a different date.

Plans and pricing units

Not yet structured in this Kingy record.

Where it runs

Availability
Not yet reviewed

Stack and integrations

Model / provider dependencies
Not yet reviewed
Integrations
Not yet reviewed

Data and regions

Vendor data-use claim
Not yet reviewed
Vendor retention claim
Not yet reviewed
Regions
Not yet reviewed

Launch and latest material update

Launch
Not yet reviewed
Latest material update
Not yet reviewed
Technical evidence and revision history
Embed “Facts tracked by Kingy”

This label is not a security certification, audit, or product endorsement. It reports the facts Kingy currently supports and when they were last checked.

Full Model Notes

Which plan is worth considering

US/USD Kimi API: US$3/M input; US$15/M output; cache writes extra. Pay-as-you-go API; minimum US$1 top-up unlocks K3. Cached input is US$0.30/M tokens; cache writes cost US$3/M for 5 minutes or US$6/M for 1 hour. Taxes extra.

Limits that affect the purchase

Thinking is always on. Rate limits depend on account tier. Self-hosting this 2.8-trillion-parameter model needs substantial infrastructure; available weights do not mean a laptop-sized download or a permissive standard license. The quoted API requires a paid top-up. Downloadable weights have no listed per-token license fee, but hosting costs and the custom Kimi K3 License still apply. This is the international USD Kimi API tariff, separate from Kimi membership, Kimi Code and OpenRouter. Current docs specify low/high/max effort, fixed sampling settings and base64 or uploaded-file references for image input.

What the existing demonstration shows

Kingy’s September 2 hosted study completed 24/24 Kimi K3 requests: mean score 94.6, 87.5% meeting the 80-point threshold, and 15.3-second median latency. That study used OpenRouter, medium effort and a 4,096-token completion cap across 12 tasks twice, with five providers involved. Total recorded cost was US$1.2126, or US$0.0577 per acceptable result. These are dated route-specific measurements, not a new native-API test or a promise about long tasks. See the retained Kingy demonstration.

Sources and buying-review date

Buying details checked October 22, 2026. Plan research uses current official sources; demonstration descriptions use the retained Kingy record and any captured creator chapters. No new subscription, paid generation or checkout was performed. The buying-review date does not renew every claim in the broader facts ledger.

Check official plans · Compare related workflows

API pricing notes

International Kimi API, USD/M tokens: input $3; output $15; cached input $0.30; cache write $3 (5 min) or $6 (1 hour).

License notes

Custom Kimi K3 License; preserve notices. Separate terms can apply to large Model-as-a-Service businesses; some large commercial products have attribution requirements. Read the exact license and exceptions.

Official Model Links

Official model page Announcement Docs Model card Pricing License Weights/download

Model Intelligence Research Map

Use these internal paths to move from this model profile into provider pages, static comparison pages, related Kingy records, and the broader AI launch graph. These are research paths, not rankings.

AI Model Intelligence Hub Compare AI Models AI Tools AI Companies AI Launches Alternatives / comparison Related article
Browse AI models Compare AI models Submit a model launch