Back to Models

Qwen: Qwen3.8 Flash

Qwentext
8.8 / 10 Overall Rating

Qwen3.8 Flash is a multimodal reasoning model from Alibaba supporting text, image, and video inputs. It features a 1,000,000 token context window and a 131,072 max output token limit at $0.15 per million input tokens. It is designed for codebase analysis, high-throughput agentic tasks, and video processing.

Context Window
1000K
Knowledge Cutoff
2025
Max Output
16K
License
Proprietary

Where it Excels (Pros)

  • 1M token context window
  • 131K max output tokens
  • Extremely cheap API pricing
  • Native video and image inputs

Limitations (Cons)

  • Weaker reasoning than Claude 3.5 Sonnet
  • Text-only output generation

Benchmark Breakdown

Reasoning (MMLU)82%
Coding (HumanEval)78%
Mathematics (MATH)76%
Long-Context Retrieval95%

Pricing Matrix

Input cost / 1M tokens$0.15
Output cost / 1M tokens$0.47

Compare Qwen: Qwen3.8 Flash with other frontier models:

Related Skills, Tools & Automations