Bench

Read news on Bench with our app.

Read more in the app

SWE-bench Verified no longer measures frontier coding capabilities

N-Day-Bench – Can LLMs find real vulnerabilities in real codebases?

Many SWE-bench-Passing PRs would not be merged

MiniMax M2.5 released: 80.2% in SWE-bench Verified

Mistral releases Devstral2 and Mistral Vibe CLI

Top model scores may be skewed by Git history leaks in SWE-bench

Qodo CLI agent scores 71.2% on SWE-bench Verified

New #1 open-source AI Agent on SWE-bench Verified

New York-focused VC Work-Bench has raised a fresh $160M

Vending-Bench: A Benchmark for Long-Term Coherence of Autonomous Agents

This Bench Does Not Exist

Bench is charging people for services they already paid for, some customers say

Bench burned through $135 million before shutting down

Failed fintech startup Bench racked up over $65 million in debt, documents reveal

Why some former Bench customers are not happy

Employer.com offers to buy failed fintech Level, a week after it makes offer for Bench

Inside the wild fall and last-minute revival of Bench, the VC-backed accounting startup that imploded over the holidays

Bench saved by the bell, and other last-minute deals that closed 2024

Bench to be acquired after abruptly shutting down

Bench shuts down, leaving thousands of businesses without access to accounting and tax docs