The Practical AI Benchmark

Fieldboard

Rules fixed in advance. Positions appear only when a season has enough matched cases.

Season 1

Real-life tasks

0/10matched cases to rank

No ranked season has been run yet.

Rules, fixed in advance

01Only scored challenges count. Perspectives never earn points.
02Models are compared only on cases they all ran.
03Points = checks passed − 2 per critical failure.
04Equal points share a position. Speed and cost are shown separately.

Round 1 · Pilot

Unranked reference

Five workplace tasks, one attempt each. Too small to rank, so no positions are given.

–GPT-6 Astra0 critical · 5 cases22/22checks
–Claude Sonnet 50 critical · 5 cases22/22checks
–Gemini 4 Argon0 critical · 5 cases21/22checks

#	Model	Checks passed	Cases	Points
–	GPT-6 AstraOpenAI	22/22	5	–
–	Claude Sonnet 5Anthropic	22/22	5	–
–	Gemini 4 ArgonGoogle	21/22	5	–

All product names, logos and brands are property of their respective owners. They are used on this site for identification and comparative benchmarking only; their use does not imply affiliation with or endorsement by them.

Benchmark briefing

How to read this benchmark

A short email guide: what the benchmark measures, how to check any result yourself, and how corrections work.