Qwen 3.8 Max Prime

Qwen 3.8 Max Prime (qwen/qwen3.8-max-prime) — the boosted tier of Qwen 3.8 Max: deeper in math, code and multilingual tasks. Reasoning depth runs from minimal to xhigh (no max); thinking is always on. Accepts images; files — no.

Parameters and capabilities

Parameter	Value
model	qwen/qwen3.8-max-prime (short alias qwen3.8-max-prime)
reasoning_effort	minimal, low, medium, high, xhigh
Disabling reasoning	no — the model always thinks
Images on input	yes (image_url, data URLs)
Files on input	no — a file part returns 400
Cache read	$0.63 per 1M tokens

Example

curl https://api.mixen.ai/v1/chat/completions \
  -H "Authorization: Bearer $MIXEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen/qwen3.8-max-prime",
    "reasoning_effort": "xhigh",
    "messages": [{"role": "user", "content": "Rewrite this SQL to an optimal plan and explain the win"}]
  }'

Pricing

Per 1M tokens: input $5, output $15, cache read $0.625. Current prices — catalog.

Strengths and limitations

• The widest depth ladder: minimal through xhigh — fine-grained spend control.
• Reasoning is always on; there is no max — the ceiling is xhigh.
• Files are not accepted on input; images are.

All models — in the catalog. The general workflow — chat guide.