DeepSeek R1 (Jan '25) vs Qwen3.7 Max
Which is better for the money: DeepSeek R1 (Jan '25) by DeepSeek or Qwen3.7 Max by Alibaba? 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
Qwen3.7 Max scores 18.1 points higher (29.5 vs 11.4) but costs 4.9× as much to run the benchmark ($2277 vs $461; flagship vs mid-range cost classes). They are not alternatives at one budget: pick Qwen3.7 Max when you need intelligence above 11.4, and DeepSeek R1 (Jan '25) when 11.4 is enough.
| Metric | DeepSeek R1 (Jan '25) | Qwen3.7 Max | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 11.4 | 29.5 | +158.8% Qwen3.7 Max is 2.6× more intelligent |
| Total AA cost (lower is better) | $460.74 | $2,277.12 | +394.2% Qwen3.7 Max is 4.9× more expensive |
| End-to-end time (lower is better) | 19.94s | 16.63s | −16.6% Qwen3.7 Max is 16.6% faster |
| Price per 1M tokens (in / out) | $2.00 / $4.00 | $2.50 / $7.50 | |
| Input modalities | Text | Text |
Qwen3.7 Max vs DeepSeek R1 (Jan '25): 158.8% more intelligence, 394.2% more cost to run the AA benchmark, and 16.6% faster end-to-end.
Today's Pareto frontier
Against the frontier
- DeepSeek R1 (Jan '25) is off the frontier: at the same or lower cost, GPT-6.1 Sol (Medium) scores 47.80 for $361.
- Qwen3.7 Max is off the frontier: at the same or lower cost, Claude Opus 5.5 (High, Default Fallback) scores 53.60 for $2172.
DeepSeek R1 (Jan '25)Qwen3.7 MaxToday'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/deepseek-r1-0120/qwen3-7-max · Markdown: /api/v1/compare/deepseek-r1-0120/qwen3-7-max?format=md · llms.txt · agent skill