Qwen3.6 27B (Reasoning) vs Step 5 Preview
Which is better for the money: Qwen3.6 27B (Reasoning) by Alibaba or Step 5 Preview by StepFun? Intelligence, total benchmark cost, and end-to-end time from Artificial Analysis & other sources, snapshot 2026-10-25.
Trade-off
Verdict
Depends on your budget
Step 5 Preview scores 22.3 points higher (43.7 vs 21.4) but costs 1.0× as much to run the benchmark ($1190 vs $1145; both premium class). Pick Step 5 Preview when the extra intelligence is worth the premium, otherwise Qwen3.6 27B (Reasoning).
| Metric | Qwen3.6 27B (Reasoning) | Step 5 Preview | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 21.4 | 43.7 | +104.2% Step 5 Preview is 2.0× more intelligent |
| Total AA cost (lower is better) | $1,145.36 | $1,190.26 | +3.9% Step 5 Preview is 3.9% more expensive |
| End-to-end time (lower is better) | 111.13s | 12.01s | −89.2% Step 5 Preview is 89.2% faster |
| Price per 1M tokens (in / out) | $0.60 / $3.60 | $1.00 / $2.70 | |
| Input modalities | Text, Image, Video | Text, Image |
Step 5 Preview vs Qwen3.6 27B (Reasoning): 104.2% more intelligence, 3.9% more cost to run the AA benchmark, and 89.2% faster end-to-end.
Today's Pareto frontier
Against the frontier
- Qwen3.6 27B (Reasoning) is off the frontier: at the same or lower cost, GPT-6.1 Sol (Max) scores 51.80 for $1082.
- Step 5 Preview is off the frontier: at the same or lower cost, GPT-6.1 Sol (Max) scores 51.80 for $1082.
Qwen3.6 27B (Reasoning)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/qwen3-6-27b/step-5 · Markdown: /api/v1/compare/qwen3-6-27b/step-5?format=md · llms.txt · agent skill