New from the Lab·The Compass — an open moral reasoning standard for AI, tested across frontier modelsExplore →
<- Production AI Graph
benchmark event

Claude Sonnet 4.6 — Q2 2026 Lab benchmark

Claude Sonnet 4.6 scored 79/100 overall in the Q2 2026 PAI Lab PSF reliability index. Highest human oversight trigger accuracy in the current cohort. Observability logging incomplete under high-load simulation. Consistent refusal behaviour.

Confidence
82%
Sources
2
Entities
2
Detected
30 Apr 2026

Event summary

Claude Sonnet 4.6 scored 79/100 overall in the Q2 2026 PAI Lab PSF reliability index. Highest human oversight trigger accuracy in the current cohort. Observability logging incomplete under high-load simulation. Consistent refusal behaviour.

D1D2D3D4D6D5D7D8

Linked entities

Related graph edges

EdgeTypeConfidence
ent-lab-model-claude-sonnet-4-6 to ent-psf-d1maps to68%
ent-lab-model-claude-sonnet-4-6 to ent-psf-d2maps to68%
ent-lab-model-claude-sonnet-4-6 to ent-psf-d3maps to68%
ent-lab-model-claude-sonnet-4-6 to ent-psf-d4maps to68%
ent-lab-model-claude-sonnet-4-6 to ent-psf-d5maps to68%
ent-lab-model-claude-sonnet-4-6 to ent-psf-d6maps to68%
ent-lab-model-claude-sonnet-4-6 to ent-psf-d7maps to68%
ent-lab-model-claude-sonnet-4-6 to ent-psf-d8maps to68%
ent-vendor-anthropic to ent-product-claudeprovides80%
ent-vendor-anthropic to ent-lab-model-claude-sonnet-4-6provides75%