Dataset

Metric scope

Headline metrics are projected; workflow evidence stays tied to analyzed samples

Projected full population996 sessions996 source objects discovered. Spend, token, and savings totals are projected to this population.

Analyzed sample300 sessionsThreads actually normalized and analyzed for workflows, findings, and evidence.

Full analysis sample300 sessionsPreview read 300 threads first.

Projection factor3.3xProjection confidence: medium. Workflow tables below remain sample-scoped.

Monthly spend · projected from 300 analyzed sessions to 996 source-pool sessions

$17.6/mo

Projected total tokens

3.8Mtokens

avg 3.8K per session

Input 802K (21%)Output 3M (79%)

Total savings available

$4.78/mo

Could cut spend by ~27%

Opportunities

Projection

3.3x

Workflows affected

View top workflow

All workflows

Token and spend trend

19 hour buckets

Dec 4 4AM

Dec 4 5AM

Dec 4 6AM

Dec 4 7AM

Dec 4 8AM

Dec 4 9AM

Dec 4 10AM

Dec 4 11AM

Dec 4 12PM

Dec 4 1PM

Dec 4 11PM

Dec 5 12AM

Dec 17 4PM

Dec 17 5PM

Dec 17 6PM

Dec 17 7PM

Dec 17 8PM

Dec 17 9PM

Dec 17 10PM

Input tokensOutput tokens

Model mix

Tokens and spend by model

6 models · all workflows

Tokens by model

input + output

Claude Opus 4.5654.2K tokens

1227 calls30.7% of total

gemini-3-pro-preview574.6K tokens

1508 calls27% of total

GPT-5.2420.6K tokens

1168 calls19.8% of total

Kimi-K2384.3K tokens

1232 calls18.1% of total

GPT-5.1-CODEX70.7K tokens

216 calls3.3% of total

Qwen3-Coder24.3K tokens

89 calls1.1% of total

Spend by model

estimated cost

Claude Opus 4.5$12.9

71.5% of total

GPT-5.2$3.39

18.9% of total

gemini-3-pro-preview$0.67

3.7% of total

Kimi-K2$0.59

3.3% of total

GPT-5.1-CODEX$0.46

2.5% of total

Qwen3-Coder$0.01

0.1% of total

Analyzed sample

Spend by workflow

7 workflows · sample-scoped

Rows below show the analyzed sample. The headline spend, tokens, and savings above are projected to the full source pool.

WorkflowRuns / moTokens (in / out)SpendSavingsTop issue

Flight Reservation Modificationderived-flight-reservation-modification10979.2K / 376.6K$2.15$0.60Tool misuse Flight Cancellation Refundderived-flight-cancellation-refund6837.9K / 139.1K$0.75$0.19Tool misuse Otherderived-other4037.5K / 104.4K$0.61$0.16Tool misuse Cabin Class Upgrade Downgradederived-cabin-class-upgrade-downgrade2929.1K / 82.6K$0.60$0.16Tool misuse New Flight Bookingderived-new-flight-booking2012.7K / 59.8K$0.47$0.14Tool misuse Booking Conflict Resolutionderived-booking-conflict-resolution1528.9K / 66.8K$0.39$0.11Tool misuse Flight Delay Compensationderived-flight-delay-compensation1916.3K / 69.4K$0.33$0.08Tool misuse

Opportunities

What's causing waste

14 opportunities · $4.78/mo projected

Tool misuse

Half of all runs fail after spending full cost, suggesting missing early checks.

$1.447 workflows

Spend by agent

7 agents

derived-flight-reservation-modification40.6%

1 workflow

derived-flight-cancellation-refund14.2%

1 workflow

derived-other11.5%

1 workflow

derived-cabin-class-upgrade-downgrade11.3%

1 workflow

derived-new-flight-booking8.9%

1 workflow

derived-booking-conflict-resolution7.4%

1 workflow

derived-flight-delay-compensation6.2%

1 workflow