MiMo-V2.6-Pro vs Qwen3.8 Max 0902: which is cheaper?

MiMo-V2.6-Pro from Xiaomi and Qwen3.8 Max (0902) from Alibaba are both tracked in Franklin AI's benchmark catalogue. The clearest gap is on the bill: Qwen3.8 Max 0902 is 5.5 times the blended price of MiMo-V2.6-Pro. The figures come from tracked catalogue data, not from our own testing, so treat them as a starting point for a trial on your tasks.

Cost: MiMo-V2.6-Pro is cheaper

Per million tokens, MiMo-V2.6-Pro charges $0.435 for input and $0.87 for output, while Qwen3.8 Max 0902 charges $2.00 and $6.00. That makes MiMo-V2.6-Pro's output 86% cheaper.

A reference job of 1 million input tokens and 200,000 output tokens costs $0.61 with MiMo-V2.6-Pro and $3.20 with Qwen3.8 Max 0902, so MiMo-V2.6-Pro saves $2.59. For a chat app with 50,000 sessions of 800 input and 300 output tokens, MiMo-V2.6-Pro would come in 82% lower: $30.45 for MiMo-V2.6-Pro against $170 for Qwen3.8 Max 0902. On a 3:1 blend of input to output tokens, that works out to $0.544 versus $3.00 per million.

Qwen3.8 Max 0902 leads on the intelligence index

Qwen3.8 Max 0902 comes out ahead on the intelligence index, 33.7 to 29.9. A lead of this size is hard to dismiss as noise. In relative terms, Qwen3.8 Max 0902 scores 13% higher.

Across the 109 active models with an intelligence score in the catalogue, MiMo-V2.6-Pro ranks 55th and Qwen3.8 Max 0902 ranks 44th.

For coding or maths work specifically, the catalogue has no dedicated index on either side yet.

Speed and latency

Once a response is under way, MiMo-V2.6-Pro produces roughly 41 tokens per second against 38 tokens per second for Qwen3.8 Max 0902. Throughput is close to level. The wait for the first token averages 3.0 seconds on MiMo-V2.6-Pro against 3.2 seconds on Qwen3.8 Max 0902.

Add the two numbers up for a 1,000-token reply and MiMo-V2.6-Pro finishes first: 27 seconds for MiMo-V2.6-Pro, 29 seconds for Qwen3.8 Max 0902.

Release timing

The release dates are September 2, 2026 for Qwen3.8 Max 0902 and September 21, 2026 for MiMo-V2.6-Pro, 19 days apart. That is close enough to count them as the same generation.

Is the price in proportion?

Dividing the blended price by the intelligence index gives a rough cost per point: $0.018 for MiMo-V2.6-Pro and $0.089 for Qwen3.8 Max 0902, per million tokens. Qwen3.8 Max 0902 costs 4.9 times as much per point, which is the price of its lead.