Alibaba vs 
Sapiens AI
Verdict
- Alibaba's flagship leads: Qwen3.8 Max at 63.3 vs Agnes 2.5 Pro Beta at 59.1.
- Alibaba has the better model in 4 of 4 price bands.
- Alibaba's fastest ranked model (Qwen3 Coder 30B A3B Instruct, 244 tok/s) beats Sapiens AI's (Agnes 2.5 Pro Beta, 113 tok/s).
- Alibaba publishes more open-weights models (48 vs 0).
Best model per price band
Each provider’s highest-index model below the price ceiling. Purple marks the winner of the band.
| Price band | Alibaba | Sapiens AI |
|---|---|---|
| Under $0.50/M | Qwen3.8-Flash-Next 60.3$0.230/M | Agnes 2.5 Pro Beta 59.1$0.150/M |
| Under $2/M | Qwen3.8-Flash-Next 60.3$0.230/M | Agnes 2.5 Pro Beta 59.1$0.150/M |
| Under $10/M | Qwen3.8 Max max 63.3$3.00/M | Agnes 2.5 Pro Beta 59.1$0.150/M |
| Any price | Qwen3.8 Max max 63.3$3.00/M | Agnes 2.5 Pro Beta 59.1$0.150/M |
Line-up
| Metric | Alibaba | Sapiens AI |
|---|---|---|
| Flagship index | 63.3 | 59.1 |
| Ranked models | 54 | 1 |
| Open-weights models | 48 | 0 |
| Cheapest ranked, $/M | $0.060 | $0.150 |
| Fastest ranked, tok/s | 244 | 113 |
| Release cadence, days | 7 | – |
| Last release | 17 Sept 2026 | – |
| Next expected | 21 Sept 2026 – 26 Sept 2026 | – |
Data as of 25 Oct 2026. See Alibaba and Sapiens AI in full.