Qwen 3.8 Max vs DeepSeek-V4-Pro
Qwen 3.8 Max reads images and video and is built to operate software; DeepSeek-V4-Pro reads text only but is a top coder under the MIT licence. Both are open, both are huge, and both sets of scores are their makers’ own.
Choose Qwen 3.8 Max if
- Agents that operate desktop software — 86.1% on OSWorld-Verified.
- Images and video as input.
- Research tasks — 93.0% on PaperBench.
Choose DeepSeek-V4-Pro if
- Pure coding and maths.
- A no-thinking mode for quick answers.
- The simplest licence: MIT.
Side by side
| Qwen 3.8 Max | DeepSeek-V4-Pro | |
|---|---|---|
| Made by | Alibaba | DeepSeek |
| Released | August 2026 | April 2026 |
| Input price, per 1M tokens | S$2.87 | S$2.03 |
| Output price, per 1M tokens | S$8.62 | S$4.06 |
| Cached input, per 1M tokens | S$0.3802 | S$0.3042 |
| Batch (about half price) | No | No |
| Context window | 1M tokens | 1M tokens |
| Longest answer | 131K tokens | 384K tokens |
| Reads | Text, Images, Video | Text |
| Thinks before answering | Yes, adjustable | Yes, adjustable |
| Open weights | Yes, Qwen3.8-Max License | Yes, MIT |
Prices are Roar AI’s published rates, live from our price list. Specs are each lab’s own published figures. Prices are our published US dollar rates converted at S$1.30 to the dollar, our rate for October — the same rate your card is charged.
What a real job costs
Take 1,000 requests, each about 3,000 tokens in and 1,000 out — a long prompt and a page of answer.
DeepSeek-V4-Pro does the same job for 1.7× less. Models that think before answering spend extra output tokens doing it, so on hard prompts the real gap can be wider than the rates suggest.
Use both, switch any time
On Roar AI both are on the same key and the same monthly invoice, paid in your local currency. Trying the other one is a one-word change:
client.chat.completions.create(model="qwen3.8-max", …) client.chat.completions.create(model="deepseek-v4-pro", …)
Send the same prompts to both for a day and compare the answers on your own work — that settles it faster than any benchmark. Read the docs.
Try both
on one key
Send the same prompts to each and compare the answers on your own work.