Alibaba
Qwen3.8 2.4T A95B
Quality Score60.9
Value Score44.3
Reliability52.2
Cache Discount88%
Model specification
- Reasoning
- reasoning
- Input modalities
- text
- Output modalities
- text
- Context window
- 983,616 tokens
- Weights
- open weights
Token prices USD per 1M tokens
- Input
- $2.00
- Output
- $6.00
- Cached Input
- $0.250
- Cached Output
- not available
- Cost per task
- $1.09
Capability
- Intelligence
- 57.7
- Coding
- 71.9
- Agentic
- 57.1
- Omniscience
- 4.3
- Correct
- 31.3%
- Blended price
- $1.18
Answer outcomes
Correct 31.3%Incorrect 26.9%Abstained 41.8%
Artificial Analysis benchmarks
- GDPval-AA v2
- 61%
- τ³-Banking
- 49%
- Terminal-Bench v2.1
- 82%
- SciCode
- 52%
- Humanity’s Last Exam
- 42%
- GPQA Diamond
- 94%
- CritPt
- 20%
- AA-Omniscience
- 52%
- AA-LCR
- 75%
Similar models
- Qwen3.8 MaxAlibabaQuality Score 61.2Value Score 44.6
- Qwen3.8 27B (xhigh)AlibabaQuality Score 47.2Value Score 41.5
- Qwen3.7 PlusAlibabaQuality Score 43.5Value Score 36.5
- Qwen3.8 27B (medium)AlibabaQuality Score 38.6Value Score 36.3
- Qwen3.8 27B (low)AlibabaQuality Score 38.5Value Score 35.1
- Qwen3.6 27B (Reasoning)AlibabaQuality Score 36.1Value Score 29.4