Intellectual hub of the topic

ai benchmarks

AI: Events

SWE-fficiency: Evaluating Not Just an AI's Bug-Finding Ability, But the Efficiency of Its Fixes

Development

A new benchmark assesses how quickly and accurately AI agents fix code, not just identify problems – taking into account time, attempts, and real-world working conditions.

OpenHandsopenhands.dev Feb 17, 2026

AI: Events

How2Everything: When Chatbot Instructions Actually Need to Work

Development

Allen AI has released a framework to evaluate and improve language models' ability to generate step-by-step instructions that actually help get the job done.

Ai2allenai.org Feb 12, 2026

AI: Events

SenseTime Unveils SenseNova-SI-1.3: A Model Featuring Advanced Spatial Intelligence

Products

The Chinese firm has open-sourced an AI model that topped the charts in eight spatial environment understanding benchmarks.

SenseTimewww.sensetime.com Feb 7, 2026

AI: Events

Perplexity Introduces Benchmark for Evaluating Deep AI Research Quality

Development

The new DRACO benchmark evaluates how accurately, thoroughly, and objectively AI systems handle complex topic exploration across various fields of knowledge.

Perplexity AIresearch.perplexity.ai Feb 6, 2026

AI: Events

Why AI Voice Agents Are Switching to Direct Speech Processing

Products

We explore how direct speech processing differs from text-based intermediate steps and why it matters for the voice assistants of the future.

Ultravoxwww.ultravox.ai Feb 3, 2026

AI: Events

OpenHands Index: A New Way to Compare AI Agents on Real-World Tasks

Development

OpenHands has launched a benchmark demonstrating how models handle real-world GitHub tasks – from bug fixes to implementing new features in open-source projects.

OpenHandsopenhands.dev Jan 29, 2026

AI: Events

Moonshot Releases Kimi K2.5 – A Model With Enhanced Reasoning and Long-Context Support

Products

Chinese company Moonshot AI has unveiled an updated version of its language model, Kimi K2.5, featuring strengthened reasoning capabilities and the ability to process documents up to one million tokens in length.

MoonShot AIwww.moonshot.ai Jan 27, 2026

AI: Events

How to Evaluate Language Models' Understanding of the Emirati Arabic Dialect

Research

Researchers from the UAE have created a benchmark suite to test how well large language models handle the Emirati Arabic dialect.

Hugging Facehuggingface.co Jan 27, 2026

AI: Events

Anthropic Unveils Economic Index to Assess AI Impact on Real-World Work

Anthropic has introduced a novel method for assessing AI progress through an economic lens, evaluating which real-world tasks models can already perform to augment or replace human efforts.

Anthropicwww.anthropic.com Jan 15, 2026

Don’t miss a single experiment!

Subscribe to our Telegram channel —
we regularly post announcements of new books, articles, and interviews.