Tatum

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

$2,485

159 sessions · $16 each

Savings opportunity

$770

31% of spend, at list price

Practices on target

3 of 7

Focus: rtk coverage, Knowledge MCP use

Cache hit rate

79%

Team 84% · org 82%

FY2027 to date · estimated at list price

$770 of $2,485 spend

Largest saving first

Improve prompt caching$435
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.
Right-size models$248
43% of routine spend ran on Opus: code research $159 of $359, documentation $113 of $265, ops $318 of $752, presentation $10 of $29.
Run docs, decks, ops and read-only research on Sonnet instead of Opus.
Keep context lean$72
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$15
63% of 1,562 Bash calls ran through rtk; 2.3M 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 Infrastructure and the organization

Practice	Tatum	Team	Org	Target	Recommendation
rtk coverage	63%Needs attention	69%	56%	≥ 80% of Bash calls	Install the rtk hook so test, build, git and search commands return compact output.
Knowledge MCP use 4 findings saved back	28%Needs attention	37%	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	84%	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	6%	10%	≤ 10% of input	On target
Opus on routine work	43%Needs attention	33%	44%	≤ 25% of routine spend	Default routine sessions to Sonnet and keep Opus for harness, design and hard debugging work.
Sessions using subagents	52%On target	51%	50%	≥ 30% of sessions	On target
Tool error rate	3.2%On target	3.1%	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	–	–	58%Off target	9%On target	62%Off target	42%On target	$49
Mar 2026	–	–	62%Off target	7%On target	62%Off target	24%Needs attention	$71
Apr 2026	–	–	66%Off target	8%On target	62%Off target	34%On target	$86
May 2026	24%Off target	–	72%Needs attention	10%On target	55%Off target	40%On target	$106
Jun 2026	49%Needs attention	10%Off target	70%Needs attention	6%On target	53%Off target	57%On target	$148
Jul 2026	74%Needs attention	22%Needs attention	74%Needs attention	9%On target	40%Needs attention	60%On target	$130
Aug 2026	75%Needs attention	32%Needs attention	79%Needs attention	11%Needs attention	38%Needs attention	55%On target	$61
Sep 2026	66%Needs attention	31%Needs attention	82%Needs attention	8%On target	30%Needs attention	61%On target	$82
Oct 2026	70%Needs attention	38%Needs attention	87%On target	7%On target	27%Needs attention	54%On target	$38

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	2,640	12M	126.7M	8.9M	3.4M	$187
Opus 5	2,807	22.9M	134.5M	10M	3.9M	$342
Opus 4.8	2,480	33.5M	110.3M	9.2M	3.8M	$375
Opus 4.6	2,947	46.1M	115.5M	10.3M	4.9M	$475
Opus 4.5	128	2.5M	4.1M	425K	202K	$22
Sonnet 5.5	4,319	22.2M	256.4M	17.8M	6.9M	$183
Sonnet 5	6,165	52.6M	293.8M	22.1M	9M	$463
Sonnet 4.6	3,367	46.6M	131.6M	11.4M	4.9M	$295
Sonnet 4.5	141	2.6M	4.2M	429K	156K	$13
Haiku 5.5	6,034	21.4M	268.4M	18.5M	9.4M	$12
Haiku 4.5	4,122	43.1M	193.4M	15.1M	6.2M	$112
GPT Realtime (meetings)	56	592K	2.1M	172K	76K	$5
Total	35,206	306.1M	1641.1M	124.3M	52.8M	$2,485

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.