Anthropic · Released 2026-10-07
Frequent smaller tasks: examine an independent score, drawing effort levels and the 100k-token price boundary.
Original model explanation ↗43points
Intelligence Index
Rank 12 among included models
Test configuration: max with fallback
Source read · · v4.3.2
Each model uses its highest tested setting on the source page, with fallback conditions retained; scores come from one evaluator. The index mainly evaluates English text tasks; examine Chinese and image capabilities separately.
Source evaluation & rankings ↗Independent evaluation
Anthropic positions Haiku 5.5 around summaries, classification, extraction and subagent work: small operations repeated many times.
Public creator test
Simon Willison tested five effort levels with a cycling-pelican prompt.
Vendor customer account and suggested use
Anthropic quotes Rogo describing a larger model preparing a deck while a Haiku subagent retrieves segment revenue from an annual report.
An update to complex work and communication, with public code-animation and same-prompt game examples.
Original release ↗The second 5.5-family model emphasizes everyday work, speed and effort-dependent costs.
Original release ↗The lightweight model gains adjustable effort and lower short-prompt rates for extraction, classification and subagent work.
Original release ↗