Tokscale
Start Now
Menu

Kimi K3 (k3)

Observed through

Tokscale Model Report

Kimi K3 (k3) model details and observed coding usage from Tokscale.

Browse models

Reality Snapshot

Observed tokens

41.2B>+999%

over 30 days

Estimated tracked spend

$15.8K

Invoices may differ

Active developers

72+71

this month

30-day momentum

>+999%

Compared with the prior 30-day window

Most-used client

Kimi48%

of observed tokens

Observed Usage

Tokens observedDaily active developers
Observed coding tokens over 30 daysDaily observed usage from Jul 12 through Aug 10, with active developer activity for context.5B4B3B2B1B050403020100Jul 12Jul 15Jul 18Jul 21Jul 24Jul 27Jul 29Aug 1Aug 4Aug 7Aug 10

How developers use it

Usage by coding client

Sharevs 30 days
Kimi
48%0%
OpenCode
20%0%
Pi
11%0%
Senpi
10%0%
Gajae-Code
8%0%
Claude Code
2%0%
Hermes Agent
<1%0%

Token composition

Share of observed tokens

97%cache read
Cache read39.8B97%
Input1.3B3.1%
Output159.8M<1%
Reasoning3.5M<1%
Cache-heavy workloadMost observed tokens reuse cached context across requests.

Observed Variants

defaultk341.2B$15.8K$0.3899.95%
via kimikimi/k311.6M$0$0.000.03%
highk3-high6.9M$0$0.000.02%
maxk3-max2.4M$0$0.000.01%

Token Economics

Estimated spend profile

Kimi$8.7K54.79%
OpenCode$3.4K21.66%
Pi$1.9K12.08%
Senpi$1.7K10.85%
Gajae-Code$0.000%
Claude Code$250.16%
Hermes Agent$710.45%
Codex CLI$2.650.02%

Observed cost signals

Estimated spend per 1M tracked tokens$0.38
Cache-read share96.5%
Estimated cache savingsNot available
Typical workloadcontext-heavy

Estimated tracked spend calculated from Tokscale model and provider pricing data. Actual invoices may differ.

Model Momentum

>+999%30-day growthRank: Observed model

Compared with the prior non-overlapping window

Fastest-growing client: Kimi

6B4B2B0Jul 12Jul 27Aug 10

Official Model Details

Tokscale Insight

Kimi K3 (k3) recorded 41.2B coding tokens across 72 developers during the latest 30-day Tokscale window.

Observed confidenceHigh