Intelligence Index
0127.8
> 74% OF MODELS
Qwen
Qwen3.8 27B
Updated: September 2026
Open-weight dense Qwen vision-language model for coding, research, professional workflows, and long-running agents.
Specifications
- Context
- 262K
- Max Output
- 131K
- Input Modalities
- Image, Text, Video
- Output Modalities
- Text
- Input
- $0.45/M
- Output
- $3.2/M
Capabilities
VISIONTHINKINGCODINGWRITING
Similarly Priced Models
| Model | Provider | Context | Input Price | Output Price |
|---|---|---|---|---|
| Kimi K2.5 | Moonshot | 262K | $0.45/M | $2.2/M |
| Qwen3 235B A22B | Qwen | 131K | $0.455/M | $1.82/M |
| DeepSeek V4 Pro | DeepSeek | 1M | $0.435/M | $0.87/M |
| DeepSeek V4 Pro | DeepSeek | 1M | $0.435/M | $0.87/M |
| DeepSeek V4 Pro | DeepSeek | 1M | $0.435/M | $0.87/M |
Performance Metrics
Coding Index
0256.1
> 62% OF MODELS
Agentic Index
0345.3
> 87% OF MODELS
Average Response Performance
- Output Speed
- 46.6 tok/s
- Time To First Token
- 3.94s
- Time To First Answer Token
- 46.84s
- End To End Response Time
- 57.57s
DATA SOURCE: Artificial Analysis