Qwen3.8 Max (0902) vs Step 5 Preview
Which is better for the money: Qwen3.8 Max (0902) by Alibaba or Step 5 Preview by StepFun? Intelligence, total benchmark cost, and end-to-end time from Artificial Analysis & other sources, snapshot 2026-10-24.
Trade-off
Verdict
Depends on your budget
Qwen3.8 Max (0902) scores 1.7 points higher (45.4 vs 43.7) but costs 4.1× as much to run the benchmark ($4935 vs $1190; flagship vs premium cost classes). They are not alternatives at one budget: pick Qwen3.8 Max (0902) when you need intelligence above 43.7, and Step 5 Preview when 43.7 is enough.
| Metric | Qwen3.8 Max (0902) | Step 5 Preview | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 45.4 | 43.7 | −3.7% Step 5 Preview is 3.7% less intelligent |
| Total AA cost (lower is better) | $4,934.79 | $1,190.26 | −75.9% Step 5 Preview is 75.9% cheaper |
| End-to-end time (lower is better) | 63.23s | 12.06s | −80.9% Step 5 Preview is 80.9% faster |
| Price per 1M tokens (in / out) | $2.00 / $6.00 | $1.00 / $2.70 | |
| Input modalities | Text, Image | Text, Image |
Step 5 Preview vs Qwen3.8 Max (0902): 3.7% less intelligence, 75.9% less cost to run the AA benchmark, and 80.9% faster end-to-end.
Today's Pareto frontier
Against the frontier
- Qwen3.8 Max (0902) is off the frontier: at the same or lower cost, Claude Opus 5.5 (Xhigh, Default Fallback) scores 56.00 for $4057.
- Step 5 Preview is off the frontier: at the same or lower cost, GPT-6.1 Sol (Max) scores 51.80 for $1082.
Qwen3.8 Max (0902)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/qwen3-8-max/step-5 · Markdown: /api/v1/compare/qwen3-8-max/step-5?format=md · llms.txt · agent skill