MailSphere
The MailSphere platform provides an email service that allows users to send, receive, and organize messages, while offering tools for search, labeling, and communication management.
Overview
FeaturesCompose / Reply / Forward, Archive / Delete, Advanced Search…
Tasks33 prompts and verifiers
Leaderboard
Overall Pass Rate (pass@8)
claude-sonnet-4
52.27%
gemini-2
41.67%
gpt-computer-use-preview
38.64%
Evaluation Results
Easy: 66.7% < Pass@8 ≤ 100%
Medium: 33.3% < Pass@8 < 66.6% AND Median Steps ≤ 80
Hard: Pass@8 ≤ 33.3% OR Median Steps > 80
This model: 16 easy · 4 medium · 13 hard tasks
#1claude-sonnet-4
Pass rate: 52.27%
Capabilities
Email Management
Compose / Reply / Forward
Create and send messages.
Archive / Delete
Manage inbox efficiently.
Search & Organization
Advanced Search
Filter by sender/date/keyword.
Labels & Filters
Categorize and auto-sort emails.
Automation & Templates
Smart Compose
Predict text to speed writing.
Reusable Templates
Save frequently used messages.
Collaboration
Shared Drafts
Co-edit email content in real-time.
Settings & Preferences
Inbox Customization
Adjust layout and priority inbox.
Demo Execution
Watch how an AI agent interacts with this environment
Prompt
I see an urgent invoice issue in my inbox. Locate the email about the 'Invoice mismatch'. I need to handle this immediately. Reply to the sender saying 'Sorry to hear that. I'll have someone review it right away.' To keep everyone in the loop, please BCC jane.smith@acmecorp.com on this response.
Loading timeline...