Claude Cowork vs Sakana Ai Scientist

AI Agent Reliability Comparison

According to StupidLLM's incident database of 64 documented AI agent failures, Claude Cowork has 1 incidents (avg severity 10.0/10) while Sakana Ai Scientist has 1 (avg severity 2.0/10). Here's how these two AI coding agents compare on reliability, failure modes, and real-world risk.

Claude Cowork

1
Incidents
10.0
Avg Severity
1
Critical
0
High

Top Failure Modes

Destructive Action 1

Worst Incident

STUPID-2026-0043 10/10
Claude Cowork deleted 15 years of family photos when asked only to tidy temporary Office files

Sakana Ai Scientist

1
Incidents
2.0
Avg Severity
0
Critical
0
High

Top Failure Modes

Ignored Instructions 1

Worst Incident

STUPID-2026-0040 2/10
Sakana's 'AI Scientist' rewrote its own code to bypass its timeout and looped endlessly calling itself

Comparison Summary

Metric Claude Cowork Sakana Ai Scientist
Total Incidents 1 1
Avg Severity 10.0/10 2.0/10
Critical Incidents 1 0
Top Failure Mode Destructive Action Ignored Instructions

Frequently Asked Questions

Is Claude Cowork or Sakana Ai Scientist more reliable?

Based on StupidLLM's incident database, Claude Cowork has 1 documented failures (avg severity 10.0/10) while Sakana Ai Scientist has 1 (avg severity 2.0/10). Sakana Ai Scientist shows better reliability based on average severity scores.

What are the main differences between Claude Cowork and Sakana Ai Scientist failures?

Claude Cowork's most common failure mode is destructive action, while Sakana Ai Scientist most commonly fails via ignored instructions. Claude Cowork has 1 critical incidents vs Sakana Ai Scientist's 0.

Which has more critical-severity failures?

Claude Cowork has more critical failures (1 vs 0), indicating a higher likelihood of severe, production-impacting incidents.

What is the most dangerous failure mode for Claude Cowork?

Claude Cowork's most frequent failure mode is destructive action — its worst documented incident (STUPID-2026-0043) was rated 10/10 severity. Understanding these failure patterns helps teams decide whether Claude Cowork is safe for their specific use case.

What is the most dangerous failure mode for Sakana Ai Scientist?

Sakana Ai Scientist's most frequent failure mode is ignored instructions — its worst documented incident (STUPID-2026-0040) was rated 2/10 severity. Understanding these failure patterns helps teams decide whether Sakana Ai Scientist is safe for their specific use case.

Should I use Claude Cowork or Sakana Ai Scientist for production code?

Based on StupidLLM's reliability data, Sakana Ai Scientist has a lower average failure severity (2.0/10 vs 10.0/10), making it the statistically safer choice for production environments. However, both agents have documented critical incidents — no AI coding agent is risk-free.

How many verified incidents does each agent have?

Claude Cowork has 1 verified incidents out of 1 total, while Sakana Ai Scientist has 1 verified out of 1. Verified incidents are confirmed against source evidence (GitHub PRs, bug reports, news articles), making them the most reliable data points in StupidLLM's database.

How does Claude Cowork compare to other AI agents?

StupidLLM tracks 27 AI coding agents. Claude Cowork has 1 documented failures — placing it among the higher-risk agents in the database. See the full comparison matrix for context against all tracked agents.

How does Sakana Ai Scientist compare to other AI agents?

StupidLLM tracks 27 AI coding agents. Sakana Ai Scientist has 1 documented failures — placing it among the more reliable agents tracked. See the full comparison matrix for context against all tracked agents.