Benchmarks

Performance results, comparisons, and evaluations of AI models across tasks, datasets, and real-world applications.

Alex Dudov Alex Dudov Benchmarks
7 months ago

AA-Omniscience Benchmark: Claude Opus 4.5 Scores 56 in Python, Gemini 3 Pro Leads JavaScript with 56

The AA-Omniscience benchmark reveals no single AI model dominates software engineering tasks, with Claude Opus 4.5 leading Python and Go while Gemini 3 Pro Preview takes JavaScript and Swift.

Eseandre Mordi Eseandre Mordi Benchmarks
7 months ago

Grokipedia Traffic Explodes 9,900% in 2 Months, Challenging Wikipedia's Dominance

Grokipedia's meteoric rise from 35,000 to 3.5 million daily visitors in just two months marks one of the fastest growth spurts in online encyclopedia history, positioning the platform as a serious challenger to Wikipedia's long-standing dominance.

Usman Salis Usman Salis Benchmarks
7 months ago

Two AI Models Built Working CLI Tools in Under 15 Minutes—No Human Help Required

GLM 4.7 and MiniMax M2.1 autonomously created command-line task runners in 10-14 minutes, handling complex coding tasks that typically take senior developers 1-2 days to complete.

Saad Ullah Saad Ullah Benchmarks
7 months ago

Video-Thinker AI Crushes 7B-Scale Models Across Multiple Benchmarks

Video-Thinker is a new multimodal AI that processes video content through internal reasoning rather than external tools. It's outperforming every other 7B-parameter video model out there across major benchmarks.

Marina Lyubimova Marina Lyubimova Benchmarks
7 months ago

Grokipedia Publishes 400,000 Articles Daily vs Wikipedia's 500—AI Speed Gap Exposed

A new comparison reveals Grokipedia adding 400,000 articles daily while Wikipedia manages around 500, showcasing the dramatic speed difference between AI-driven and human-curated knowledge publishing.

Marina Lyubimova Marina Lyubimova Benchmarks
7 months ago

QwenLong-L1.5 Handles 4 Million Tokens, Matches GPT-5 on Long-Context Tests

QwenLong-L1.5 can reason across documents containing up to 4 million tokens—equivalent to about 100 novels—matching top long-context AI models like GPT-5 on benchmark evaluations.

Victoria Bazir Victoria Bazir Benchmarks
7 months ago

NVDA-Linked AI Benchmarks: Falcon-H1R-7B Scores 16 on Intelligence Index Among 7B Models

Technology Innovation Institute's Falcon-H1R-7B, a 7-billion-parameter reasoning model, scored 16 on the Artificial Analysis Intelligence Index v4.0, landing on the Pareto frontier for sub-12B models and marking UAE's growing presence in the AI benchmarking landscape.

Usman Salis Usman Salis Benchmarks
7 months ago

DiffThinker AI Model Beats GPT-5 by 314% in Visual Reasoning Tasks

A fresh research paper unveils DiffThinker, a diffusion-powered AI system that tackles visual reasoning challenges by creating images rather than text. Researchers say it crushed GPT-5 and Gemini-3-Flash across seven benchmark tests.

Sergey Diakov Sergey Diakov Benchmarks
7 months ago

Claude 4.5 Opus Thinking Scores 1960 Rating, Surpassing Earlier Models by Wide Margin

New benchmark ratings reveal Anthropic's Claude 4.5 Opus Thinking model achieving a 1960 rating—implying over 99% theoretical win-rate versus Claude 3 Opus—as AI model evaluation systems move toward standardized comparison frameworks.

Saad Ullah Saad Ullah Benchmarks
7 months ago

China's IQuest-Coder Beats GPT-5.1 and Claude 4.5 with 81.4% SWE-Bench Score

China's new open-source coding model IQuest-Coder has outperformed GPT-5.1 and Claude Sonnet 4.5 on several major coding benchmarks, hitting 81.4% on SWE-Bench Verified and 81.1% on LiveCodeBench v6—all while running on just 40 billion parameters.

Eseandre Mordi Eseandre Mordi Benchmarks
7 months ago

MiniMax M2.1 Coding Model Scores 88.6% on Benchmark, Beats Claude and Gemini

MiniMax just dropped its M2.1 coding model as open-source, and it's making waves by beating both Claude Sonnet 4.5 and Gemini 3 Pro on multilingual coding tests. The real eye-opener? An 88.6% score on MiniMax's VIBE full-stack benchmark.

Peter Smith Peter Smith Benchmarks
7 months ago

Nvidia GB200 Delivers 15× Better Performance Per Dollar in AI Computing

Fresh benchmark data reveals Nvidia's GB200 NVL72 cuts AI token costs dramatically compared to rivals, proving that raw performance matters more than sticker price in modern AI infrastructure.

Saad Ullah Saad Ullah Benchmarks
8 months ago

ChatGPT Crushes Gemini with 67.6M Daily Users vs 3.8M

Latest data reveals ChatGPT's massive advantage in daily active users over Google Gemini, though market dynamics show shifting competitive landscape in AI platforms.

Saad Ullah Saad Ullah Benchmarks
8 months ago

Step-DeepResearch Scores 61.4% on AI Benchmark, Rivaling Google and OpenAI at Lower Cost

StepFun has released Step-DeepResearch, a deep-research AI agent that scores 61.4% on the Scale AI Research Rubrics benchmark with a 32B model, positioning it near OpenAI and Alphabet's Gemini DeepResearch while emphasizing lower cost.

Eseandre Mordi Eseandre Mordi Benchmarks
8 months ago

Gemini and Grok Overtake ChatGPT in Average Visit Time

New data reveals Gemini.com beat ChatGPT.com in average session time during September, while Grok.com pulled ahead in October and has held the lead for two months straight.

Saad Ullah Saad Ullah Benchmarks
8 months ago

Meta AI Surpasses Claude with 5 Million Daily Active Users in Late 2024

Meta AI overtook Claude in daily active users during late 2024, powered by rapid adoption growth. Similarweb data reveals Meta AI ended the year with significantly higher engagement across mobile platforms worldwide.

Eseandre Mordi Eseandre Mordi Benchmarks
8 months ago

8B AI Model Achieves 94.5% Score, Beating GPT-5 with New PaCoRe Framework

Fresh research unveils PaCoRe, a parallel reasoning system that smartly scales computing power during testing. Using this approach, an 8-billion parameter model outscored GPT-5 on a challenging math benchmark.

Eseandre Mordi Eseandre Mordi Benchmarks
8 months ago

MiniMax M2.1 vs GLM-4.7: 74.0 vs 73.8 on SWE-bench

First benchmark results put MiniMax M2.1 and GLM-4.7 neck-and-neck on SWE-bench Verified, though MiniMax pulls ahead on Terminal Bench 2.0. Keep in mind these are vendor-reported numbers, not independent tests.

Eseandre Mordi Eseandre Mordi Benchmarks
8 months ago

GPT-5 vs GPT-4 Chart Sparks Heated Debate: Are We Actually Racing Toward AGI?

A viral radar chart comparing GPT-5 and GPT-4 reveals massive jumps in language, math, and reasoning abilities, fueling claims that AI has already blown past human experts in key intellectual areas.

Eseandre Mordi Eseandre Mordi Benchmarks
8 months ago

Alibaba's $5 Image Generator Beats Competition with 1,160 Elo Score

Alibaba's Z-Image Turbo has topped the Artificial Analysis Image Arena leaderboard, securing first place among open-weights text-to-image models with significantly lower pricing than competitors.

Posts feed
OpenAI's $122 Billion Superapp Strategy Signals a Major AI Platform Shift
4 months ago
BitVLA Shrinks Robot AI by 11x While Boosting Speed 4.4x
4 months ago
LFM2.5-350M Debuts With Strong Benchmark Results Across 6 Evaluation Suites
4 months ago
Grok 4.20 Multi-Agent Beta Ranks #7 in Search Arena Leaderboard
4 months ago
Claude Code Leak Reveals 4 Key Features Powering Autonomous AI Systems
4 months ago
RISE Robot Framework Boosts Task Success Rates by 35% Using AI-Driven Imagination
4 months ago
Multi-Agent AI Systems Face 15 Emerging Social Risks, New Study Finds
4 months ago
Mistral AI Raises $830M to Build Paris AI Data Center
4 months ago
OpenAI Codex Adds Claude Code CLI Integration in Cross-Platform Move
4 months ago
AI Coding Boom: Claude Code NPM Downloads Surge 70x, Codex Installs Hit 5M+
4 months ago
Follow us on Twitter

Aigazine is an independent publication covering artificial intelligence — new models, research, the companies building them, and the industries they reshape. Published from Tbilisi, Georgia (0179, Tbilisi City, Vake District, 49 Besarion Zhghenti Street, VAT 305786600).

© 2026 Aigazine

We use cookies to improve your experience on our site and to show you relevant advertising. To find our more, read our privacy policy and cookie policy