Capabilities · Concept
ORCA-bench: Oncall Root-Cause Analysis for Coding Agents
Production-fidelity benchmark where agents diagnose incidents from live telemetry, logs, traces, and code.
Forecast horizon
Exposes a large gap between coding skill and real SRE-style incident reasoning.
Connections
Connections · 3
How this node ties into the rest of the map, and the evidence behind each link.
to · informs 4/10
Narrowing Cyber Capability Gap of Open-Weight Models Low RCA accuracy of frontier coding agents under realistic oncall noise complements cyber-capability gap measurements for operational risk.
+4 growth to · constrains 3/10
AI Agents Reshaping Knowledge Work Low oncall RCA accuracy shows coding agents remain far from reliable production incident knowledge work.
+3 growth to · applies to 3/10
AI Agents Reshaping Knowledge Work Oncall RCA benchmark shows frontier coding agents still fail most realistic production incident tasks.
+3 growthSignal sources
Signal sources
Dated facts from primary sources in this direction.
Task horizon doubling Mar 2025
The length of software tasks AI agents can do autonomously at 50% reliability has doubled about every 7 months — and since 2024 closer to every ~3 months.
METR → Benchmarks saturating Apr 2025
In one year scores rose by 18.8, 48.9 and 67.3 points on MMMU, GPQA and SWE-bench; real-world software solve rate jumped from 4.4% to 71.7%.
Stanford HAI — AI Index 2025 → Autonomous coding 2025–2026
On SWE-bench Verified (500 real GitHub issues), autonomous coding agents reached ~80–86% by late 2025, up from under 50% in early 2025.
Epoch AI →