リリース2026/08
生成速度37 トークン/秒
初動レイテンシ (TTFT)1.82s
入力料金$2/1M
出力料金$6/1M
ランキング
レビュー
Qwen3.8 2.4T A95B excels in reasoning and coding, making it well suited for computationally intensive logic and software development tasks. Deployments must factor in cost and latency, particularly if the primary use case is general chat or autonomous agents.
長所
- Exceptional reasoning capabilities, ranking among the top 10 on AIM
- Strong programming skills, well suited for code generation and debugging
- Massive 2.4T scale providing deep contextual comprehension and broad multilingual coverage
短所
- General chat is not industry-leading, with occasional unnatural phrasing
- AI agent performance is limited and cannot fully replace specialized autonomous systems
- High compute requirements result in larger deployment costs than lighter models
ユースケース
Complex data analysis and logical deduction for research applicationsAutomated code generation, optimization, and bug fixing for software projectsDrafting technical documentation and reports requiring deep reasoning and broad context