Blake

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

$2,030

184 sessions · $11 each

Savings opportunity

$654

32% of spend, at list price

Practices on target

3 of 7

Focus: Opus on routine work, rtk coverage

Cache hit rate

77%

Team 82% · org 82%

FY2027 to date · estimated at list price

$654 of $2,030 spend

Largest saving first

Improve prompt caching$358
Cache hit rate 77% 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$212
88% of routine spend ran on Opus: code research $313 of $351, documentation $146 of $166, ops $27 of $31, presentation $27 of $31.
Run docs, decks, ops and read-only research on Sonnet instead of Opus.
Keep context lean$62
8% 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$23
60% of 1,575 Bash calls ran through rtk; 2.6M 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 Content Pipeline and the organization

Practice	Blake	Team	Org	Target	Recommendation
rtk coverage	60%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 6 findings saved back	29%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	77%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	8%On target	9%	10%	≤ 10% of input	On target
Opus on routine work	88%Off target	57%	44%	≤ 25% of routine spend	Default routine sessions to Sonnet and keep Opus for harness, design and hard debugging work.
Sessions using subagents	53%On target	55%	50%	≥ 30% of sessions	On target
Tool error rate	2.8%On target	2.4%	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	–	–	57%Off target	7%On target	91%Off target	31%On target	$60
Mar 2026	–	–	64%Off target	9%On target	–	42%On target	$51
Apr 2026	–	–	69%Off target	11%Needs attention	91%Off target	54%On target	$65
May 2026	26%Off target	–	70%Needs attention	9%On target	90%Off target	35%On target	$77
Jun 2026	50%Needs attention	12%Off target	73%Needs attention	7%On target	90%Off target	45%On target	$127
Jul 2026	72%Needs attention	25%Needs attention	74%Needs attention	6%On target	88%Off target	53%On target	$106
Aug 2026	65%Needs attention	33%Needs attention	82%Needs attention	7%On target	88%Off target	56%On target	$55
Sep 2026	66%Needs attention	38%Needs attention	81%Needs attention	10%On target	86%Off target	57%On target	$82
Oct 2026	68%Needs attention	34%Needs attention	87%On target	9%On target	86%Off target	71%On target	$31

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	3,032	17M	160.7M	11.3M	5.1M	$259
Opus 5	4,609	34.2M	221M	16.3M	6.2M	$539
Opus 4.8	3,209	39.5M	147.5M	11.9M	4.5M	$459
Opus 4.6	3,744	57.4M	147.5M	13.1M	5.3M	$576
Opus 4.5	151	3.3M	5.1M	536K	220K	$28
Sonnet 5.5	445	2.5M	25.6M	1.8M	710K	$19
Sonnet 5	955	7.6M	44.7M	3.3M	1.3M	$69
Sonnet 4.6	525	7.7M	23M	2M	841K	$50
Sonnet 4.5	31	720K	1.1M	118K	49K	$4
Haiku 5.5	578	2.5M	28.8M	2M	729K	$1
Haiku 4.5	553	6.1M	26.5M	2.1M	903K	$16
GPT Realtime (meetings)	108	1.2M	4.9M	390K	148K	$11
Total	17,940	179.7M	836.6M	64.9M	26.1M	$2,030

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.