Opus 5.5 has a serious contender on Salesforce work.
Gemini 4 Argon vs Claude Opus 5.5, 272 real tasks, default effort:
Within 3 tries: 96% vs 95%
First try: 90% vs 90%
Cost with retries: $1.68 vs $15.43
Statistically a tie, at a ninth of the price.
forcebench.ai/leaderboard/
After a decade and a half working in the Salesforce ecosystem, I launched @forcebench: an open benchmark of AI models on Salesforce work: 272 tasks in Apex, Flow, LWC, SOQL, packaging and more, graded by hidden tests. Apex and Flow run in real scratch orgs.
2.0M Followers 2 FollowingClaude is an AI assistant built by @anthropicai to be safe, accurate, and secure. Talk to Claude on https://t.co/ZhTwG8d1e5 or download the app.