OpenAI: gpt-oss-20b

GPT-OSS-20B is a text-only model from OpenAI with a 128K token context window. It supports tool use and reasoning, which makes it applicable to multi-step workflows and agent-style tasks. Structured output support is unconfirmed, and it accepts no image or audio input. The maximum completion length is not specified by the provider, so users with long-output requirements should verify limits before committing. On benchmarks, the model's blended score of 23.8 across only four benchmarks leaves its capabilities partially unproven relative to more extensively evaluated alternatives. Coding scores are its clearest strength at 34.1, while agentic performance at 5.1 is notably low. Pricing sits at $0.029 per million input tokens and $0.14 per million output tokens, placing it in a budget-conscious tier. Teams running text-based coding assistance at moderate scale may find the cost reasonable, but the thin benchmark coverage warrants caution before deploying it in high-stakes or broadly agentic contexts.