Scoper Index
The Scoper Index
A monthly read on where agents are useful in business work.
Live record May 2026
Work reviewed
0+
real jobs checked
Work types
0+
common business tasks
Findings kept
0+
useful enough to share
Checks run
0+
quality checks completed
Cost table
Model costs.
Public model docs show the running cost, working room, and image price. Prices are shown in USD.
Claude Opus 4.7
1M tokens
$5.00 per 1M
Claude Sonnet 4.6
1M tokens
$3.00 per 1M
GPT-5.5
under 270K tokens
$5.00 per 1M
Grok 4.3
1M tokens
$1.25 per 1M
ChatGPT Image
$0.034
GPT Image 1.5 medium, 1024px
Grok Imagine
$0.02
per image
Grok Imagine Quality
$0.05
per image
Source check: OpenAI, Anthropic, and xAI model docs, 14 May 2026.
Index view
Prepared tasks.
Agents are useful for first drafts, routine admin, organised reads, and simple visuals. Important decisions still need a person.
Index score
68/100
Higher means the work is easier to use.
Work strength
Useful today
May 2026
Provider scores
Claude, OpenAI, Grok
May 2026
Working room
Running cost
Connected tools
Saved instructions
Specialist info
Higher scores mean the provider is stronger for that category.
Working room
A higher score means the provider can hold more context while completing the task.
Useful for long briefs, large files, and multi-step work.
Running cost
A higher score means repeat work is cheaper to run at useful quality.
Useful when the same job happens often.
Connected tools
A higher score means stronger support for connectors, MCP, and business apps.
Useful when the agent needs to work across systems.
Saved instructions
A higher score means repeat instructions are easier to reuse.
Useful when a workflow should run the same way next time.
Specialist info
A higher score means stronger support for specialised business knowledge.
Useful for legal, medical, finance, and regulated work.
Index scale
Four levels.
Ready
Useful now with user review.
Useful
Worth using as a starting point.
Review
Needs careful checking before use.
User-led
The agent prepares. The user decides.