Together AI: gpt-oss-120B vs o4-mini — Real-World Performance Comparison

AI Impact Summary

OpenAI has released gpt-oss models, including a 20B parameter reasoning model and a 120B parameter model, designed to rival o4-mini. Together AI conducted a comparative analysis across five practical tests, demonstrating gpt-oss-120B’s superior performance in code generation, creative tasks, complex instruction following, mathematical reasoning, and information synthesis. These results highlight the potential of open-source models for enterprise use cases, particularly in scenarios requiring customization and cost-effectiveness.