Which AI models can you automate on Zapier? (GPT-5.6 Sol, Gemini 3.6 Flash, Opus 5, and more)
Zapier has updated its AutomationBench leaderboard to rank AI models by performance in real-world business workflows, highlighting OpenAI’s GPT-5.6 Sol, Anthropic’s Opus 5, and Google’s Gemini 3.6 Flash as top performers across different use cases.
The useful question is what changes for users, developers or buyers, and whether the announcement stays industry context or becomes something people can actually use.
Zapier’s AutomationBench evaluates AI models on multi-step business workflows rather than static prompts, measuring their ability to handle complex, real-world tasks. The benchmark includes scenarios like resolving scheduling conflicts by applying priority policies, rescheduling events, and posting updates to Slack. Models are scored solely on whether they complete the task correctly, regardless of the tools or steps used. This approach ensures workflows are tested under conditions that mimic actual business environments, including irrelevant data and ambiguous instructions.
OpenAI’s GPT-5.6 family—Sol, Terra, and Luna—are now available on Zapier, each optimized for different workflows. GPT-5.6 Sol (Max) ranks second in Sales with a 16.0% completion rate, while Terra and Luna target multi-tool coordination and high-volume, rule-based tasks, respectively. OpenAI’s lineup spans budget-friendly mini models to advanced reasoning engines, with some older models being deprecated from the API over time.
Anthropic’s Claude models, particularly Opus 5, lead across most business functions on the AutomationBench leaderboard. Opus 5 (Max) achieves a 26.5% completion rate in Sales and a 56.0% rate in Operations, outperforming other models in Support, Marketing, Finance, and HR. Known for strong writing quality and safety-first design, Claude models are well-suited for drafting content, analyzing documents, and powering customer-facing chatbots with natural conversational tone.
Google’s Gemini 3.6 Flash (High) is now available on Zapier, excelling in step coordination and strict policy adherence at a low cost of $0.58 per task. It ranks sixth overall with a 19.8% completion rate, making it a cost-effective choice for high-volume workflows. However, it may struggle with strict output formats and independent mathematical decision-making. The Gemini family is ideal for processing long documents and research-heavy tasks where cost efficiency is a priority.