Tokscale
Open console
Menu

K3

Observed through

Tokscale Model Report

K3 model details and observed coding usage from Tokscale.

Official details Browse models

Reality Snapshot

Observed tokens

16.7B

over 30 days

Estimated tracked spend

$6.1K

Invoices may differ

Active developers

53

this month

30-day momentum

—

Awaiting a prior 30-day window

Most-used client

Kimi65%

of observed tokens

Observed Usage

Tokens observedDaily active developers
Observed coding tokens over 30 daysDaily observed usage from Jun 30 through Jul 29, with active developer activity for context.5B4B3B2B1B02520151050Jun 30Jul 3Jul 6Jul 9Jul 12Jul 15Jul 17Jul 20Jul 23Jul 26Jul 29

How developers use it

Usage by coding client

Sharevs 30 days
Kimi
65%0%
OpenCode
12%0%
Pi
10%0%
Gajae-Code
8%0%
Claude Code
5%0%
Hermes Agent
<1%0%

Token composition

Share of observed tokens

96%cache read
Cache read16B96%
Input606.3M3.6%
Output73.5M<1%
Reasoning948.3K<1%
Cache-heavy workloadMost observed tokens reuse cached context across requests.

Token Economics

Estimated spend profile

Kimi$4.5K73.27%
OpenCode$80913.22%
Pi$77112.6%
Gajae-Code$0.000%
Claude Code$250.4%
Hermes Agent$300.5%
Codex CLI$0.640.01%

Observed cost signals

Estimated spend per 1M tracked tokens$0.37
Cache-read share95.9%
Estimated cache savingsNot available
Typical workloadcontext-heavy

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

Model Momentum

—30-day growthRank: Observed model

Awaiting a prior non-overlapping window

Fastest-growing client: Kimi

4B2.67B1.33B0Jun 30Jul 15Jul 29

Official Model Details

Tokscale Insight

K3 recorded 16.7B coding tokens across 53 developers during the latest 30-day Tokscale window.

Observed confidenceHigh