AI Benchmarks

What is AICrier?

Live AI developer news, ranked and linked to original sources.

> ▌

BENCHMARK×
⌘K
THURSDAY // 2026-11-19
25 items
NOV 19
Anthropic's Still Winning...
PT24M59S
// WATCH
YouTube

Anthropic's Still Winning...

+1

Ben Davis

Google Unleashes Gemini 4 Neon, Microsoft Drops 3 New Models, New RSI, Kling Reaches 8K...
PT13M9S
// WATCH
YouTube

Google Unleashes Gemini 4 Neon, Microsoft Drops 3 New Models, New RSI, Kling Reaches 8K...

+7

AI Revolution

Gemini 4 Neon vs Carbon: Which One Should You Actually Use?
PT8M7S
// WATCH
YouTube

Gemini 4 Neon vs Carbon: Which One Should You Actually Use?

+1

DIY Smart Code

Genie 4 Can Summon a Spaceship. Will It Stay There?
PT3M
// WATCH
YouTube

Genie 4 Can Summon a Spaceship. Will It Stay There?

DIY Smart Code

I'm gonna miss coding
PT28M16S
// WATCH
YouTube

I'm gonna miss coding

Theo - t3․gg

Claude Code Animations That Don't Look Generic (No Other Subs)
PT9M50S
// WATCH
YouTube

Claude Code Animations That Don't Look Generic (No Other Subs)

+1

Eric Michaud

NVIDIA Finally Fixed Claude Skills By Releasing This
PT12M51S
// WATCH
YouTube

NVIDIA Finally Fixed Claude Skills By Releasing This

+2

AI LABS

Gemini 4 Neon Is #3. Its Own Table Shows Where It Loses
PT2M58S
// WATCH
YouTube

Gemini 4 Neon Is #3. Its Own Table Shows Where It Loses

+1

DIY Smart Code

Can the Resolution Hypothesis Destroy AI?
PT39M32S
// WATCH
YouTube

Can the Resolution Hypothesis Destroy AI?

+4

Discover AI

Gemini 4 Neon Is INSANE – Is THIS Google's BEST Model Yet?
PT39M51S
// WATCH
YouTube

Gemini 4 Neon Is INSANE – Is THIS Google's BEST Model Yet?

+1

Bijan Bowen

🚨🚨 Curl Rewrite to Rust - All Joy Gone 🚨🚨
P0D
// WATCH
YouTube

🚨🚨 Curl Rewrite to Rust - All Joy Gone 🚨🚨

+1

The PrimeTime

Qwen4-Max vs Opus 5.5: 4x Cheaper Tokens, Same Bill
PT8M26S
// WATCH
YouTube

Qwen4-Max vs Opus 5.5: 4x Cheaper Tokens, Same Bill

+2

DIY Smart Code

GitHub Trending Weekly #60: rust-anything, trnrt, datacat, obscura, tamserve, zigzag, openwren
PT14M32S
// WATCH
YouTube

GitHub Trending Weekly #60: rust-anything, trnrt, datacat, obscura, tamserve, zigzag, openwren

+11

Github Awesome

Meet Brian and Beast, the dots at work at Holiday Extras
PT46S
// WATCH
YouTube

Meet Brian and Beast, the dots at work at Holiday Extras

OpenAI

Google Gemini 4 Neon vs GPT 6.2 vs Opus 5.5: 8 Websites Built Blind ($4.52)
PT9M46S
// WATCH
YouTube

Google Gemini 4 Neon vs GPT 6.2 vs Opus 5.5: 8 Websites Built Blind ($4.52)

Income stream surfers

DeepMind’s New AI Is A Gift To Humanity
PT6M26S
// WATCH
YouTube

DeepMind’s New AI Is A Gift To Humanity

Two Minute Papers

Thanks Google! I am cancelling my Claude subscription...
PT15M33S
// WATCH
YouTube

Thanks Google! I am cancelling my Claude subscription...

+1

AICodeKing

Gemini 4 Neon: Between Hydrogen and Carbon
PT1M13S
// WATCH
YouTube

Gemini 4 Neon: Between Hydrogen and Carbon

Prompt Engineering

GPT-6.2 Astra hits #2 on MacroscopeBench
BENCHMARK// 2h ago

GPT-6.2 Astra hits #2 on MacroscopeBench

OpenAI's GPT-6.2 Astra and Sol models have debuted on Macroscope's code review benchmark, which tests defect detection across real-world pull requests. Astra claimed the #2 overall spot with an 83.2 F1 score at high effort—finishing behind only Claude Opus 5.5 max effort—while Sol took #4 and achieved the leaderboard's highest precision at 91.6%.

macroscopebenchgpt-6.2benchmarkevaluationcode-reviewai-codingreasoningllm+8+7+6+5+4+3+2+1
Odysseys2 benchmark challenges frontier browser agents
BENCHMARK// 4h ago

Odysseys2 benchmark challenges frontier browser agents

Odysseys2 is a web agent benchmark developed by Carnegie Mellon University and Browser Use to evaluate AI agents across 100 complex, live-web tasks lasting up to 60 minutes. An LLM judge grades final delivered files alongside screenshots and action histories to penalize reward hacking, leaving leading models completing fewer than half the tasks.

odysseys2browser-useweb-agentagentbenchmarkevaluationcomputer-use+7+6+5+4+3+2+1
NO_SCREENSHOT
BENCHMARK// 4h ago

Gestalt tops design leaderboard with 1931 Elo

A stealth AI model codenamed Gestalt has surged to the top of Contra Labs' Frontier Design Benchmark with an unprecedented 1,931 Elo rating. Evaluated blindly by professional designers on production UI briefs, the unannounced model outperformed Claude Fable 5.2 by 218 points and cleared the human baseline.

benchmarkevaluationai-codingmultimodalllmgestalt+6+5+4+3+2+1
Gemini 4 Neon tops search leaderboard
BENCHMARK// 4h ago

Gemini 4 Neon tops search leaderboard

Google DeepMind's Gemini 4 Neon claimed the top spot on Parallel's Search Intelligence leaderboard, outperforming all rival frontier models while operating at up to ten times lower cost. The mid-tier reasoning model establishes a new Pareto frontier for autonomous web research and agentic retrieval.

gemini-4-neonllmbenchmarkevaluationweb-agentreasoningsearch+7+6+5+4+3+2+1
Gemini 4 Neon tested on blind web design
BENCHMARK// 14h ago

Gemini 4 Neon tested on blind web design

Google's Gemini 4 Neon went head-to-head against GPT-6.2 and Opus 5.5 across eight single-file HTML, CSS, and JavaScript site builds, completing the entire benchmark suite for $4.52. The test demonstrates that frontier models can reliably generate functional, styled front-end prototypes in single prompts at negligible API cost.

gemini-4-neongeminibenchmarkevaluationcode-generationai-coding+6+5+4+3+2+1
Gemini 4 Neon takes #2 on HearingBench
BENCHMARK// 13h ago

Gemini 4 Neon takes #2 on HearingBench

Google's mid-tier reasoning model reached the second spot on Adam Holter's HearingBench leaderboard, trailing only Gemini 4 Carbon. The benchmark challenges vision-language models to identify sounds, words, and full sentences purely from visual spectrogram images without audio input.

gemini-4-neonmultimodalvisionbenchmarkevaluationreasoningllm+7+6+5+4+3+2+1
Gemini 4 Neon tops Coding Agent Index
BENCHMARK// 14h ago

Gemini 4 Neon tops Coding Agent Index

Google's Gemini 4 Neon has overtaken Anthropic's Claude Fable 5.2 on the Artificial Analysis Coding Agent Index, scoring 59 at extra-high reasoning effort. The mid-tier model delivers top-tier agentic coding benchmark performance at roughly half the operational cost and higher token throughput than rival frontier models.

gemini-4-neonbenchmarkcoding-agentai-codingevaluationllmreasoning+7+6+5+4+3+2+1
WEDNESDAY // 2026-11-18
2 items
NOV 18
Google's Neon is wild...
PT14M18S
// WATCH
YouTube

Google's Neon is wild...

+1

Wes Roth

GPT 6.2 Astra is out!
PT25M39S
// WATCH
YouTube

GPT 6.2 Astra is out!

+2

AI Search