Jordan

Lead engineer · AI usage, harness practices and where spend could have gone further.

Demonstration report. This engineer and their usage are synthetic.

This page is for the engineer and their manager to find cheaper defaults and harness fixes. It is never used to rank people.

Period

Summary

FY2027 to date

AI spend

$2,407

241 sessions · $10 each

Savings opportunity

$1,005

42% of spend, at list price

Practices on target

2 of 7

Focus: Opus on routine work, rtk coverage

Cache hit rate

79%

Team 79% · org 82%

FY2027 to date · estimated at list price

$1,005 of $2,407 spend

Largest saving first

Right-size models$402
81% of routine spend ran on Opus: code research $516 of $639, documentation $305 of $380, ops $24 of $29, presentation $133 of $164.
Run docs, decks, ops and read-only research on Sonnet instead of Opus.
Improve prompt caching$358
Cache hit rate 79% against an 85% target.
Keep stable context at the front of prompts and avoid switching models mid-session, so 85% of input reads from cache.
Keep context lean$209
25% of input tokens were past 100K of context.
Compact or /clear between tasks instead of carrying more than 100K tokens of context.
Route commands through rtk$36
61% of 2,172 Bash calls ran through rtk; 3.8M tokens of command output went into context.
rtk compacts test, build, git and search output before it reaches the model (about 70% smaller).

Compared with Growth & SEO and the organization

Practice	Jordan	Team	Org	Target	Recommendation
rtk coverage	61%Needs attention	62%	56%	≥ 80% of Bash calls	Install the rtk hook so test, build, git and search commands return compact output.
Knowledge MCP use 9 findings saved back	30%Needs attention	30%	30%	≥ 40% of sessions	Search the knowledge MCP before re-researching a codebase, and save findings back with remember.
Cache hit rate	79%Needs attention	79%	82%	≥ 85% of input	Keep CLAUDE.md and tool lists stable, avoid mid-session model switches, and resume sessions instead of restarting.
Context past 100K	25%Needs attention	18%	10%	≤ 10% of input	Compact or /clear between tasks; hand broad searches to subagents so their reads stay out of the main context.
Opus on routine work	81%Off target	58%	44%	≤ 25% of routine spend	Default routine sessions to Sonnet and keep Opus for harness, design and hard debugging work.
Sessions using subagents	44%On target	47%	50%	≥ 30% of sessions	On target
Tool error rate	3.0%On target	3.0%	3.0%	≤ 4% of tool calls	On target

Is the engineer moving toward the targets?

Month	rtk	Knowledge	Cache	Context	Model fit	Subagents	Savings opportunity
Feb 2026	–	–	60%Off target	28%Off target	86%Off target	10%Off target	$92
Mar 2026	–	–	61%Off target	25%Needs attention	86%Off target	24%Needs attention	$109
Apr 2026	–	–	69%Off target	24%Needs attention	86%Off target	29%Needs attention	$64
May 2026	26%Off target	–	69%Off target	24%Needs attention	84%Off target	30%On target	$261
Jun 2026	50%Needs attention	13%Off target	77%Needs attention	23%Needs attention	83%Off target	52%On target	$100
Jul 2026	67%Needs attention	21%Needs attention	80%Needs attention	24%Needs attention	79%Off target	47%On target	$121
Aug 2026	66%Needs attention	33%Needs attention	83%Needs attention	27%Off target	78%Off target	46%On target	$109
Sep 2026	69%Needs attention	39%Needs attention	86%On target	25%Needs attention	76%Off target	57%On target	$96
Oct 2026	76%Needs attention	39%Needs attention	87%On target	23%Needs attention	75%Off target	60%On target	$54

FY2027 to date

By model tier, by month

Routine types (documentation, presentation, ops, code research) rarely need Opus

Cache reads are billed at a fraction of uncached input; a high share is what good caching looks like

Model	Requests	Uncached input	Cache reads	Cache writes	Output	Spend
Opus 5.5	3,308	14M	164.4M	11.4M	4.7M	$239
Opus 5	5,362	29.4M	251.2M	17.9M	7.6M	$574
Opus 4.8	3,481	30.5M	158.9M	12.1M	5.3M	$439
Opus 4.6	4,596	74.5M	183.4M	16.5M	7.3M	$749
Opus 4.5	175	3.5M	6.2M	621K	248K	$31
Sonnet 5.5	845	4M	48M	3.3M	1.4M	$36
Sonnet 5	2,182	13.1M	105.7M	7.6M	3.4M	$150
Sonnet 4.6	1,222	17M	51.3M	4.4M	1.8M	$110
Sonnet 4.5	79	1.7M	2.9M	292K	107K	$9
Haiku 5.5	1,806	7.9M	96.7M	6.7M	2.3M	$4
Haiku 4.5	1,619	15.8M	79.9M	6.1M	2.5M	$44
GPT Realtime (meetings)	233	2.4M	11M	854K	365K	$23
Total	24,908	213.8M	1159.8M	87.7M	36.9M	$2,407

Spend and tokens come from LiteLLM gateway logs and imported agent session history, priced at API list rates.
Savings are estimates: routine Opus work repriced at the Sonnet available at the time; uncached input moved to cache reads up to an 85% hit rate; rtk removing 70% of unfiltered command output, which stays in context for about half the session's remaining turns; and 50% of input past 100K context avoided by compacting. Levers overlap, so the total is an upper bound.
rtk and the knowledge MCP are scored from their rollout (rtk May 2026, knowledge MCP Jun 2026).
Generated Oct 20, 2026, 7:46 AM.