Alibaba has officially launched its new flagship large language model, Qwen3.8, boasting a total parameter count of 2.4 trillion. This advanced model demonstrates significant improvements in programming and professional office tasks, positioning it among the top global large models. Qwen3.8-Max, which supports visual understanding and a 1M token context length, is slated for open-source release next week, alongside the Qwen3.8-27B model. The new model offers enhanced reasoning capabilities, faster inference speeds, and competitive pricing for API services. AI
IMPACT Sets new SOTA on coding and visual reasoning benchmarks, potentially accelerating enterprise adoption of advanced AI agents.
RANK_REASON Frontier-lab model release with system card [lever_c_demoted from frontier_release: ic=2 ai=1.0]
- Alibaba
- Anthropic
- Arena
- Claude
- PaperBench
- Qwen3.8
- Qwen3.8-Max
- Vision Arena
- Agent's Last Exam
- BabyVision
- GitHub
- GPQA Diamond
- Hermes Agent
- IF Bench
- Opus5
- OSWorld-Verified
- Qwen3.8-27B
- RecreationBench
- WideSearch
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →