ProgramBench

AI

Sonnet 5.5's fastest setup used subagents; Opus 5.5's used a five-agent team in one ProgramBench evaluation

A user reports that the two models ran fastest with different setups in a multi-agent evaluation on ProgramBench.

1 Source, 15h ago, first seen 15h ago

AI

Agensh reportedly coordinates 1,024 coding agents without a central orchestrator

A post sharing a Microsoft Research paper says Agensh coordinates coding agents through shared state. Agents claim subtasks, share findings, verify results and merge changes asynchronously.

EL

JY

OP

9 Sources, 12h ago, updated 12h ago

AI

MirrorCode Tops Jack Clark Newsletter, Seeks Anthropic System Card Adoption

1 Source, 61d ago, first seen 61d ago

AI

Meta Researcher Urges Paradigm Shift in Coding Benchmarks

AGI House hosts Meta researcher Kilian Lieret for talk on agentic coding benchmarks.

2 Sources, 61d ago, first seen 61d ago

ProgramBench

Recent stories

Sonnet 5.5's fastest setup used subagents; Opus 5.5's used a five-agent team in one ProgramBench evaluation

Agensh reportedly coordinates 1,024 coding agents without a central orchestrator

MirrorCode Tops Jack Clark Newsletter, Seeks Anthropic System Card Adoption

Meta Researcher Urges Paradigm Shift in Coding Benchmarks

ProgramBench

Recent stories

Sonnet 5.5's fastest setup used subagents; Opus 5.5's used a five-agent team in one ProgramBench evaluation

Agensh reportedly coordinates 1,024 coding agents without a central orchestrator

MirrorCode Tops Jack Clark Newsletter, Seeks Anthropic System Card Adoption

Meta Researcher Urges Paradigm Shift in Coding Benchmarks