Unverified50% confidenceBenchmarkExact time
The test covered 6 repositories, 1777 SWE-bench tasks, and 5 agent-and-model combinations, with a per-run cost of about 200 USD.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
11/14/2026
First Seen
Valid until: 2/12/2027
Sources
Microsoft Open-Sources Token Economizer: Saving 18%-31% on AI Coding Agent Costs
bilibili量子菠萝_11/11/2026
Related Entities
Related Claims
UnverifiedAt the highest-tier setting on Automation Bench, Opus 5.5 achieves a task success rate of 40%, with an average model cost of approximately $1.28 per attempt.64% similarUnverifiedThe test involved a total of 66 calls and cost $4.36, including 12 voided images.64% similarUnverifiedUnder the fixed test fixture (BUDGET=200,000, spent 60,000, SAFETY=0.80, REPLAY=0.15), the planning ceiling is 112,000.64% similarUnverifiedA hands-on blog post tested RTK by spending over $1,500 and running more than 1,700 tasks.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1170780API
curl https://kongchang.com/api/v1/knowledge/claims/1170780MCP
get_claim(id=1170780)