Alibaba's Qwen3.7-Max and DeepSeek V4-Pro are both 1M-context models from Chinese labs, but DeepSeek prices far lower and varies its rate by time of day. Qwen3.7-Max is also no longer Alibaba's flagship — Qwen3.8-Max replaced it at a lower price.
DeepSeek V4-Pro is far cheaper: $0.66/$1.98 per 1M off-peak, and still under Qwen3.7-Max at DeepSeek's 2× peak rate of $1.32/$3.96. If you are set on Qwen, look at Qwen3.8-Max ($2.00/$6.00) before paying $2.50/$7.50 for the older model.
| Qwen3.7-Max | DeepSeek V4-Pro | |
|---|---|---|
| Provider | Alibaba | DeepSeek |
| Tier | Mid-tier | Mid-tier |
| Input / 1M tokens | $2.50 | $0.66 |
| Output / 1M tokens | $7.50 | $1.98 |
| Cache hit / 1M | $0.25 | $0.022 |
| Cache write / 1M | — | — |
| Context window | 1M | 1M |
| Notes | Prior Qwen flagship; strong multilingual | A planned Sep 14, 2026 reroute to V4.1-Flash rates was reversed after user pushback; billing unchanged since the Aug 13, 2026 GA build. Off-peak rate shown; peak (01:00-04:00 & 06:00-10:00 UTC weekdays) is 2x: $1.32/$3.96 |
DeepSeek V4-Pro is 3.8× cheaper on input tokens ($0.66 vs $2.50 per 1M). DeepSeek V4-Pro is 3.8× cheaper on output tokens ($1.98 vs $7.50 per 1M).
| Workload | Qwen3.7-Max | DeepSeek V4-Pro | Difference |
|---|---|---|---|
| Support chatbot 10,000 conversations × (2,000 input + 400 output tokens) | $80.00 | $21.12 | DeepSeek V4-Pro saves $58.88 |
| Document Q&A 1,000 questions × (30,000 input + 800 output tokens) | $81.00 | $21.38 | DeepSeek V4-Pro saves $59.62 |
| Coding agent session 4M input tokens (90% cache hits) + 150,000 output tokens | $3.02 | $0.64 | DeepSeek V4-Pro saves $2.38 |
Standard rates per 1M tokens from data on this site. The coding-agent row assumes 90% of input is read from the prompt cache (full input price where a model has no cache rate). Cache-write fees, batch discounts and long-context surcharges listed in the notes are not included. Updated November 2, 2026.
you already run Qwen in production, rely on the strong multilingual output it is listed for, and have a reason not to move to the cheaper Qwen3.8-Max yet.
cost per token is the priority; even at peak hours it undercuts Qwen3.7-Max, and its cache hits cost a fraction of Qwen's.
The DeepSeek rate in the table is off-peak. On weekdays from 01:00–04:00 and 06:00–10:00 UTC the price doubles to $1.32/$3.96 per 1M; weekends and Chinese public holidays are off-peak all day, according to DeepSeek's pricing page. Those windows cover the European morning and late Asian afternoon, so a European team's daytime traffic may land partly at peak. Batch jobs can simply be scheduled outside them. Qwen3.7-Max has no time-of-day pricing in our data.
V4-Pro's cache hit is $0.022 per 1M versus $0.25 on Qwen3.7-Max — more than ten times cheaper. For agents and chat apps that resend long histories, that pushes the effective input cost even further apart than the headline rates suggest, as the coding-agent row above shows.
Qwen3.7-Max is listed as the prior Qwen flagship. Qwen3.8-Max, released Aug 3, 2026, costs $2.00 input and $6.00 output with a $0.17 cache hit — lower than 3.7-Max on all three — so staying on 3.7-Max only makes sense while you re-validate prompts.
Yes, at both off-peak ($0.66/$1.98) and peak ($1.32/$3.96) rates, versus $2.50/$7.50 for Qwen3.7-Max.
01:00–04:00 and 06:00–10:00 UTC, Monday to Friday, excluding Chinese public holidays. All other times are billed at the off-peak rate.
No. Qwen3.8-Max is the newer flagship and is cheaper per token than Qwen3.7-Max.
Yes, both list 1M tokens.
How we choose: prices come from each vendor's official pricing page and are re-checked every week. We take no affiliate fees, so rankings aren't for sale.