Atlas · C10 · Core cognition
Reasoning on novel problems
Solving problems never seen before, without a template (CHC: fluid reasoning).
AI Reliable score
5 / 10
Average adult level · peak 8 · trend ↑↑ · grade B
Why this score
Fastest-moving capability of 2026: on ARC-AGI-3 (interactive, novel) the best score went from 0.37% at launch (March) to ~63% at maximum compute (October), against humans ~100%. Reliable 5 because the high scores need very high compute and spatial/abstract deduction is still weak.
On the human scale
2Beginner. Solves familiar puzzles
4Average adult. Average adult: figures out simple new problems ← AI today
6Competent pro. Professional: solves unfamiliar work problems without a template
8Expert. Expert: invents approaches to hard new problems
10Best humans. Best human problem-solvers
Future warning
At its best, AI already reaches 8 on this capability. Peak results are never counted in the index; they show where the water may rise next.
Evidence
ARC-AGI-3 (interactive novel-reasoning games)B
E13 · 2026-03 to 2026-10
- Measures
- Solving never-seen interactive puzzles
- AI result
- Launch (Mar 2026): best 0.37%. Oct 28, 2026 snapshot: GPT-6 Astra (Max) 62.7% semi-private (mirrored from ARC Prize by BenchLM).
- Humans
- Humans ~100% (every task solvable by people)
- On our scale
- Fastest-rising capability, but high scores need maximum compute. To be verified on arcprize.org.
officechai.com ↗
benchlm.ai ↗
MathSpatial-Bench (arXiv 2602.11635 v2)B
E06 · 2026-04
- Measures
- Mathematical spatial reasoning from images
- AI result
- Best model GPT-5 58.5%; most models below 60%.
- Humans
- Humans ~96.3%
- On our scale
- Spatial reasoning below average adult: pulls C05 Reliable down.
arxiv.org ↗
A Definition of AGI (Hendrycks et al., CHC-based)B
E07 · 2025-10
- Measures
- 10 CHC cognitive domains, each worth 10% of an 'AGI score'
- AI result
- GPT-5 total 57%: Knowledge 9, Reading-Writing 10, Maths 10, Reasoning 7, Working memory 4, LT memory storage 0, LT memory retrieval 4, Visual 4, Auditory 6, Speed 3 (out of 10 each).
- Humans
- Well-educated adult (definition of 100%)
- On our scale
- Used as a cross-check on the profile shape, not for absolute levels (single model, dated Oct 2025).
arxiv.org ↗
History
First reading: Atlas v0.1, 5 Nov 2026. Every later change will be listed here with its reason.
Related AI Pulse
WATCH SIGNAL · pressure rising 28 Oct 2026C10
BenchLM (mirroring ARC Prize) B · source ↗
Disagree? Send a correction
