Claude Haiku 5.5 (Low) vs Claude Sonnet 4.6 (Max)
Which is better for the money: Claude Haiku 5.5 (Low) by Anthropic or Claude Sonnet 4.6 (Max) by Anthropic? Intelligence, total benchmark cost, and end-to-end time from Artificial Analysis & other sources, snapshot 2026-10-26.
Trade-off
Verdict
Depends on your budget
Claude Sonnet 4.6 (Max) scores 0.6 points higher (30.1 vs 29.5) but costs 144× as much to run the benchmark ($4925 vs $34.3; flagship vs budget cost classes). They are not alternatives at one budget: pick Claude Sonnet 4.6 (Max) when you need intelligence above 29.5, and Claude Haiku 5.5 (Low) when 29.5 is enough.
| Metric | Claude Haiku 5.5 (Low) | Claude Sonnet 4.6 (Max) | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 29.5 | 30.1 | +2.0% Claude Sonnet 4.6 (Ma… is 2.0% more intelligent |
| Total AA cost (lower is better) | $34.31 | $4,924.80 | +14253.8% Claude Sonnet 4.6 (Ma… is 144× more expensive |
| End-to-end time (lower is better) | 11.45s | 148.96s | +1201.0% Claude Sonnet 4.6 (Ma… is 1201.0% slower |
| Price per 1M tokens (in / out) | $0.10 / $0.50 | $3.00 / $15.0 | |
| Input modalities | Text, Image | Text, Image |
Claude Sonnet 4.6 (Max) vs Claude Haiku 5.5 (Low): 2.0% more intelligence, 14253.8% more cost to run the AA benchmark, and 1201.0% slower end-to-end.
Today's Pareto frontier
Against the frontier
- Claude Haiku 5.5 (Low) is off the frontier: at the same or lower cost, GPT-6 Luna (Medium) scores 29.90 for $31.2.
- Claude Sonnet 4.6 (Max) is off the frontier: at the same or lower cost, Claude Opus 5.5 (Xhigh, Default Fallback) scores 56.00 for $4057.
Claude Haiku 5.5 (Low)Claude Sonnet 4.6 (Max)Today's Pareto frontierEvery other model
Matchup
Compare something else
Agents:
Accept: text/markdown or ?format=md returns this page as Markdown. JSON: /api/v1/compare/claude-haiku-5-5-low/claude-sonnet-4-6-adaptive · Markdown: /api/v1/compare/claude-haiku-5-5-low/claude-sonnet-4-6-adaptive?format=md · llms.txt · agent skill