Qwen 3 4B Instruct 2507
Qwen develops the Qwen 3 4B Instruct 2507, a chat model exceling at instruction following, logical reasoning, and text generation, with notably enhanced capabilities in long-tail knowledge coverage and long-context understanding, supported by its 262,144 token context window. The model demonstrates strong performance in reasoning benchmarks, such as AIME25 and ZebraLogic, and also shows high scores in knowledge benchmarks like MMLU-Pro and MMLU-Redux.
Input
Output
Context
262K
Max Output
262K
Parameters
4B
Technical Specifications
Input Modalities
Output Modalities
Capabilities
Resources & Links
Estimated GPU Requirements for Qwen 3 4B Instruct 2507
Estimates based on INT8 quantization. Actual requirements vary by framework and configuration.
Related Tools
Data sourced from official provider APIs and documentation
Last updated: Jun 24, 2026
Start building with the right model.
Automatically route workloads to the right model for every task, every time.