Benchmark Watch

Evaluation notes, benchmark interpretation, leaderboard skepticism, and measurement failures.

Deep Dives and Frameworks

Implementation playbooks, operator patterns, and durable analysis.

No deep-dive content is currently available for this path.

External tools

Execution tooling is separate

Swarm Signal keeps the analysis layer. Use BoredTools for reusable production templates and trackers.

Open BoredTools Open Budget Tracker
Tool-Use Agents Need Failure Labels, Not Pass Rates

Tool-Use Agents Need Failure Labels, Not Pass Rates

Tool-use agents can fail in ways a final accuracy score hides, because the same wrong answer can come from skipped tools, ignored outputs, fabricated...

4 min read
Computer-Use Agents Fail Long Workflows, Not Mouse Clicks

Computer-Use Agents Fail Long Workflows, Not Mouse Clicks

Computer-use agents are clearing more short benchmark tasks, but the new failure line is workflow length. A June 2026 benchmark called OSWorld 2.0 tests...

5 min read
Agent Leaderboards Can Be Cheaper Without Being Safer

Agent Leaderboards Can Be Cheaper Without Being Safer

A March 2026 paper on efficient agent benchmarking found that mid-difficulty task subsets can remove large parts of an agent benchmark while preserving...

5 min read
Multimodal Memory Tests Expose the Personal-Agent Gap

Multimodal Memory Tests Expose the Personal-Agent Gap

Product teams are turning memory into the selling point for personal agents. The hard question is no longer whether they can remember a preference; it is...

5 min read
Power Grid Agents Need Constraint Tests, Not Chat Scores

Power Grid Agents Need Constraint Tests, Not Chat Scores

A June 2026 power-systems benchmark argues that language-model agents can solve grid-engineering tasks, but the useful signal is narrower: the agent must...

5 min read
TerminalWorld Makes Agent Benchmarks Harder to Fake

TerminalWorld Makes Agent Benchmarks Harder to Fake

TerminalWorld turns public terminal recordings into validated agent tasks. The signal is not a higher leaderboard score. It is a harder benchmark supply chain.

4 min read
Million-Token Context Still Fails the Workload Test

Million-Token Context Still Fails the Workload Test

Anthropic reported on February 5, 2026 that Claude Opus 4.6 scored 76% on the 8-needle 1M-token MRCR v2 test while Claude Sonnet 4.5 scored 18.5% on the...

7 min read
Coding Agent Benchmarks Hit the Generalization Wall

Coding Agent Benchmarks Hit the Generalization Wall

Scale's SWE-Bench Pro public leaderboard reports that top models scoring above 70% on SWE-Bench Verified fall to 23.3% for OpenAI GPT-5 and 23.1% for...

6 min read
Self-Improving Agents Have an Evaluator Problem

Self-Improving Agents Have an Evaluator Problem

Anthropic's June 2026 update on recursive self-improvement is not a distant sci-fi warning. The company says its engineers now ship 8x as much code per...

3 min read
The 12-to-72 Problem: Computer-Use Agents Hit Human Scores but Miss the Point

The 12-to-72 Problem: Computer-Use Agents Hit Human Scores but Miss the Point

Computer-use agents jumped from 12% to 72% on OSWorld in 18 months. The scores look like progress. The latency and efficiency numbers tell a different story.

4 min read
Swarm Signal
0:00
0:00
Up Next

Queue is empty. Click "+ Queue" on any article to add it.