Release Date: Jun 1, 2026

Developer Alibaba 🇨🇳
Context Window 1M
Max Output Tokens 66k
Token Costs (in/out) $0.40/1.60
Weights Private
Input Modalities

Accuracy

38.65 % ± 1.17

Cost / Test (Vals Index)

$ 0.440

Latency

21 min 48 s

Vals Index
BenchmarksAccuracyRankings

0.0%

±1.17
31/46

0.0%

±2.93
36/49

0.0%

±3.00
21/22

0.0%

±3.18
29/46

0.0%

±1.04
36/49

0.0%

±2.57
31/49

0.0%

±0.93
35/95

0.0%

±3.38
48/76

0.0%

±4.61
39/84

0.0%

±4.33
11/28

0.0%

±0.65
39/54
Proprietary BenchmarksAcademic BenchmarksIndustry Partners
Vals
Default Provider : Alibaba
Temperature: 0.7
Top P: Default
Top K: Default
Max Output Tokens: 65,536

Updates

Jun 2, 2026

The model was run via the Alibaba API at temperature 0.7, with preserve_thinking enabled, a 1M-token context window, and 65,536 max output tokens.