Claude Haiku 5.5 (Medium) vs Step 5 Preview
Which is better for the money: Claude Haiku 5.5 (Medium) by Anthropic or Step 5 Preview by StepFun? Intelligence, total benchmark cost, and end-to-end time from Artificial Analysis & other sources, snapshot 2026-10-25.
Trade-off
Verdict
Depends on your budget
Step 5 Preview scores 9.2 points higher (43.7 vs 34.5) but costs 21× as much to run the benchmark ($1190 vs $56.4; premium vs mid-range cost classes). They are not alternatives at one budget: pick Step 5 Preview when you need intelligence above 34.5, and Claude Haiku 5.5 (Medium) when 34.5 is enough.
| Metric | Claude Haiku 5.5 (Medium) | Step 5 Preview | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 34.5 | 43.7 | +26.7% Step 5 Preview is 26.7% more intelligent |
| Total AA cost (lower is better) | $56.36 | $1,190.26 | +2011.9% Step 5 Preview is 21× more expensive |
| End-to-end time (lower is better) | 10.33s | 12.01s | +16.3% Step 5 Preview is 16.3% slower |
| Price per 1M tokens (in / out) | $0.10 / $0.50 | $1.00 / $2.70 | |
| Input modalities | Text, Image | Text, Image |
Step 5 Preview vs Claude Haiku 5.5 (Medium): 26.7% more intelligence, 2011.9% more cost to run the AA benchmark, and 16.3% slower end-to-end.
Today's Pareto frontier
Against the frontier
- Claude Haiku 5.5 (Medium) is on today's frontier: nothing is both smarter and cheaper.
- Step 5 Preview is off the frontier: at the same or lower cost, GPT-6.1 Sol (Max) scores 51.80 for $1082.
Claude Haiku 5.5 (Medium)Step 5 PreviewToday's Pareto frontierOpen-weights 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-medium/step-5 · Markdown: /api/v1/compare/claude-haiku-5-5-medium/step-5?format=md · llms.txt · agent skill