#benchmarking

Why We Benchmark Against Industry, Not Just Company: A Psychometrics Explainer

A candidate scores above most employees in your organisation. That sounds encouraging — until you discover the company’s existing talent…

The frontier is exactly where HLE stops telling models apart

An IRT audit of 428 HLE items finds one factor behind the eight domain scores, and precision that peaks below where the best models sit.

What Is AI Benchmarking?

What those AI scores actually mean

Evals Are the Cornerstone of Every AI solution. So Why Doesn’t Anyone Build One?

You don’t need a specialist eval platform to find out if a model works for you. You need a weekend, 28 tasks files, and one API key

How AI models claim that they are the BEST?

Every AI company says their model is the best.

I Ran the Same 50 Questions Through 3 AI Models. Here’s What Actually Happened.

A simple look at what “benchmarking” means, and why the winner on paper isn’t always the model you should use.

The Bottleneck Isn’t Me

A frontier model’s note on where the money goes, written during its seven-day release window

The Evolution of the Football Giant

SuperFast Labs — High‑Performance Algorithm Benchmark,a technical demonstration of a multi‑stage…

This article presents a technical demonstration of a multi‑stage computational benchmark developed at SuperFast Labs. The benchmark…

What Matters When Benchmarking Cyber Security in LLMs (And Why I Don’t Trust Existing Approaches)

We have an ever-growing ecosystem of AI-driven tools designed to help us catch vulnerabilities and defend our codebases. But for security…

Why Static Vector Database Benchmarks Miss Production Reality

I have stopped trusting vector database benchmarks that only measure a fully loaded, fully optimized, read-only collection.

Everyone’s Hyping Self-Evolving AI Agents — But Can We Actually Prove They Get Better?

Micro-Benchmarking Java with JMH and Bazel

Ooooh do i love benchmarking .. one of my favourite things to do at work. ( i could go on and on about parca but something else for today…)

Major, path-breaking trends and standout papers from CVPR 2026.

Based on the official proceedings and accepted papers, the conference marked a distinct shift away from traditional per-scene optimization…

Beyond Clusters and Plots: Building CellFateBench, a Benchmark for Single-Cell Genomics Reasoning

How CellFateBench turns single-cell workflows into public tasks, hidden answer keys, oracle outputs, scoring, calibration, Docker…

Prefill, Decode, and TTFT: The Three Numbers That Should Drive Your Inference Engine Choice

When teams pick an LLM inference engine, the conversation usually starts and ends with one word: throughput. How many tokens per second can…

Building a Rust Benchmarking Agent

Building a Cross Cloud Rust Agent using A2A and MCP

Benchmarking Is Not About Finding a Winner

When most people see a benchmark, they immediately look for the top-ranked product.

SurrealDB 3.x by the numbers

Author: Tobie Morgan Hitchcock

LLM Benchmarking and Evals

Why Enterprise AI Needs More Than Benchmark Scores

Benchmarking I-APL, Sharp APL/PC and SuperPET Waterloo microAPL

Let’s time the benchmarks from Dick Holt’s “A beginner’s guide to low-cost “real work” APL interpreters” using I-APL and Sharp APL/PC, and…

Endee vs Vespa: A Comprehensive Vector Database Benchmark

Vector databases are increasingly at the heart of production AI systems — powering semantic search, recommendation engines, and…

ASR Evaluation Framework: Benchmarking Speech Recognition Models Across Accuracy, Speed, and…

Picking an ASR model for production is not straightforward. Whisper might be the most accurate for general English but too slow for…

Why Graded Systems Fail CCM Vocalists (and What We Do Instead)

Stilogerizerist Series — Week 4 (Bonus)

Building a “Truth” Dataset for Bioinformatics Method Development

What do you do when the research problem you aim to explore with deep learning lacks a gold-standard dataset? You create your own; like me.

Your JS Benchmark Breaks When GC Joins In

Learn why JavaScript benchmarks lie under garbage collection pressure and how to measure performance the way production really feels it.

UKRI Lakukan Kegiatan Kunjungan Benchmarking ke Universitas Pakuan (Unpak)

“Berharap kunjungan ini dapat menjadi awal dari kerja sama yang lebih luas dan berkelanjutan antara kedua institusi.”

AI’s Big Capability Claims Depend on Who Does the Grading

When AI companies make big claims about future capabilities, financial markets move and media amplifies. But the question nobody asks is…

Benchmarking ADHD Care: 10 Metrics Midwest Clinics Track

Stop Using Python Loops

How I Achieved a 1000x Speedup with CuPy on Google Colab.

Evaluating AI Models: When Metrics Don’t Lie But Aren’t the Ultimate Truth (I)

Evaluating an artificial intelligence model requires trusting metrics. That’s not bad. Like a compass, metrics let us know our position…

FG & Multi-FG: Does it Work?

In recent years Frame Generation has become a hot topic of debate among tech-gaming reviewers and enthusiasts. Now supported by all three…

Cinebench 2026

After a long wait, Cinebench supports 2025’s GPU generation: NVIDIA’s Blackwell and AMD’s RDNA4. Coming in the final week of 2025…

10 New ChatGPT Prompts For Benchmarking Analysis

99% of people benchmark wrong.

Benchmarking KDB-X vs QuestDB, ClickHouse, TimescaleDB and InfluxDB with TSBS

Key Takeaways

New GPU, Old API: Retrobenchmarks

In this story I will test two GPUs: the RTX 5080 and the RX 9070 XT, but running “retro” benchmarks and games this time.

What Is a Function in Wolfram Mathematica, Really? Part II

Writing a Function to List all Mathematica Symbols with an Assumptions option and asking LLMs to do the same.

DOOM: Path Tracing and Bechmarks

DOOM: The Dark Ages finally brought together all my computer hobbies: rendering tech, benchmarking, and killing monsters. The game at…

Benchmarking Psikologi Umsida dengan UM Purworejo, Bentuk Kolaborasi dan Kembangkan Kurikulum!

fpip.umsida.ac.id — Fakultas Psikologi dan Ilmu Pendidikan (FPIP) Universitas Muhammadiyah Sidoarjo (Umsida) mengadakan kegiatan…

9070 XT and RTX 5080: Overclocking

This is Part IV of a series of reviews: Part I for the RTX 5080, Part I for the 9070 XT, combined Part II with Microbenchmarks, Part III…

GravityMark Revisited

One year and a half ago, and one GPU-generation ago, in Dec 2013 I published RDNA 3, GravityMark and Linux. TLDR: this new benchmark had…

How Top 1% Companies Use Benchmarking to Steal Competitors’ Clients

In business, market share isn’t given — it’s taken.  The top 1% of companies know this truth better than anyone.  They aren’t just…

Benchmarking for Startups: When to Stop Guessing and Start Measuring

Startups are built on instinct, speed, and relentless energy.  But at a certain point, instinct isn’t enough.

Benchmarking 101: Tools, Key Benefits, Features, & Trends

In today’s fast-paced and hypercompetitive landscape, businesses can’t afford to operate in a vacuum. Knowing where your organization…

9070 XT and RTX 5080: Undervolting

This is Part III of a series of reviews: Part I for the RTX 5080, Part I for the 9070 XT, combined Part II with Microbenchmarks.

UX Competitive Analysis and Benchmarking: A Guide with Examples from Pakistan’s Mobile Payment…

Introduction

AI Hardware vs. AI Software

RDNA 4 is a big milestone for AMD’s efforts in AI on client: for gaming with FSR 4.0 and more coming later, but also for general AI…

9070 XT and RTX 5080: Microbenchmarks

This replaces my previous post "RTX 5080: Microbenchmarks". I now have also a 9070 XT for testing, so I updated my charts with first-party…

RTX 5080: Games & DLSS

RTX 5080: DLSS

SciHorizon: Benchmarking AI-for-Science Readiness from Scientific Data to Large Language Models

배경 및 개요: