Emery

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

$3,627

204 sessions · $18 each

Savings opportunity

$841

23% of spend, at list price

Practices on target

3 of 7

Focus: rtk coverage, Knowledge MCP use

Cache hit rate

83%

Team 82% · org 82%

FY2027 to date · estimated at list price

$841 of $3,627 spend

Largest saving first

Improve prompt caching$444
Cache hit rate 83% 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.
Right-size models$240
43% of routine spend ran on Opus: code research $317 of $782, documentation $96 of $196, ops $136 of $313, presentation $33 of $71.
Run docs, decks, ops and read-only research on Sonnet instead of Opus.
Keep context lean$111
9% 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$46
1% of 1,675 Bash calls ran through rtk; 4.1M 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 Video Rendering and the organization

Practice	Emery	Team	Org	Target	Recommendation
rtk coverage	1%Off target	2%	56%	≥ 80% of Bash calls	Install the rtk hook so test, build, git and search commands return compact output.
Knowledge MCP use 7 findings saved back	28%Needs attention	28%	30%	≥ 40% of sessions	Search the knowledge MCP before re-researching a codebase, and save findings back with remember.
Cache hit rate	83%Needs attention	82%	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	9%On target	17%	10%	≤ 10% of input	On target
Opus on routine work	43%Needs attention	43%	44%	≤ 25% of routine spend	Default routine sessions to Sonnet and keep Opus for harness, design and hard debugging work.
Sessions using subagents	46%On target	48%	50%	≥ 30% of sessions	On target
Tool error rate	2.9%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	8%On target	62%Off target	19%Needs attention	$44
Mar 2026	–	–	62%Off target	9%On target	62%Off target	20%Needs attention	$60
Apr 2026	–	–	68%Off target	8%On target	62%Off target	31%On target	$86
May 2026	0%Off target	–	69%Off target	9%On target	55%Off target	38%On target	$101
Jun 2026	5%Off target	14%Off target	74%Needs attention	7%On target	53%Off target	53%On target	$264
Jul 2026	0%Off target	19%Needs attention	79%Needs attention	10%Needs attention	40%Needs attention	44%On target	$133
Aug 2026	0%Off target	34%Needs attention	84%Needs attention	9%On target	38%Needs attention	53%On target	$48
Sep 2026	0%Off target	37%Needs attention	86%On target	11%Needs attention	30%Needs attention	50%On target	$57
Oct 2026	0%Off target	33%Needs attention	89%On target	9%On target	27%Needs attention	57%On target	$48

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	4,526	16.7M	262.7M	17.8M	7.6M	$361
Opus 5	5,485	28.6M	258M	18.3M	7.8M	$582
Opus 4.8	5,717	59.7M	249.4M	19.7M	7.8M	$742
Opus 4.6	3,285	52M	135.1M	11.9M	5.3M	$535
Opus 4.5	86	1.9M	3.3M	329K	134K	$16
Sonnet 5.5	7,097	22.3M	378.6M	25.6M	9.9M	$245
Sonnet 5	8,106	50.6M	418.5M	29.9M	12.6M	$579
Sonnet 4.6	4,397	54.9M	179.2M	14.9M	5.8M	$362
Sonnet 4.5	119	2.5M	4.4M	436K	174K	$13
Haiku 5.5	12,314	34.8M	630.5M	42.5M	18.7M	$24
Haiku 4.5	6,785	49.2M	301.4M	22.4M	9.2M	$153
GPT Realtime (meetings)	175	1.2M	7.9M	582K	212K	$14
Total	58,092	374.3M	2829M	204.5M	85.3M	$3,627

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.