Google: Gemma 4 31B (free)

Google's Gemma 4 31B (free) accepts text, image, and video input and returns text, with a context window of 262,144 tokens and a maximum completion length of 32,768 tokens. It supports tool calling and a reasoning mode, which makes it usable for agent-style workflows and multi-step tasks. The listing does not confirm structured output support or whether the weights are open, so check those details before relying on either. The main draw is price: input and output both cost nothing, so it suits prototyping, high-volume experiments, and budget-limited projects that need long context or video input. Its benchmark evidence is thin. It has a blended score of 47.9 across only two independent benchmarks, with 24.2 on AI Index and 70.8 on EQ-Bench, so its general capability is not well established. Shortlist it as a no-cost option and test it on your own workload before committing.