Vega Research Copilot

Document Summariser

complex-agenticconfidence high — a separate axis from tier; low confidence never auto-decides
Vega Research Copilot›Knowledge Baselabel · app=knowledge-base›Document Summariserwf-vg021c9read throughAFVega Researchsub-2290-vega

Where this came from

Every figure below is only as trustworthy as the account it was read through

Application
Knowledge Base
identified by label · app=knowledge-base
Project
Vega Research Copilot
Vega Sciences
Fingerprint
wf-vg021c9
inferred from logs

Read through 1 account

Cost

30-day spend

$19,272

LOGGED

On recommended path

$18,678

MEASURED

Share of attributed spend

104%

Current → recommended (diffed against production)

In production today

gemini-2.5-pro

reasoning high · cap 8,192

Recommended

gemini-2.5-pro (unchanged)

reasoning low · cap 2,048

  • Δ reasoning: high → low
  • Δ response cap: 8,192 → 2,048

Replay evidence

Model

gemini-2.5-pro · reasoning low

Samples

120

Input tokens

1,480,000

Output tokens

61,000

Reasoning tokens

19,000

Latency

1850 ms end-to-end

Judge verdict

HOLDS

Replay spend

$6.40

MEASURED — the only grade replay can produce, and the only stage that spends money
cannot say — session identifiers do not exist in this log source