category · 225 articles · latest 10.11.2026 · page 2
Benchmarks
benchmarks·GPT-6.1 Terra vs Gemini 4 Argon: Polish Test ResultsI tested GPT-6.1 Terra and Gemini 4 Argon on Polish prompts, office tasks, technical drawings and forms.benchmarks·GPT-6.1 Terra Computer Use: Form-Filling Test
benchmarks·VEED Dubbing on fal.ai: Polish Voice Cloning TestI tested VEED Dubbing on fal.ai with Polish and English clips.
benchmarks·Gemini 4 Carbon benchmarks: all 19 tests and where it losesGoogle reports Gemini 4 Carbon leads in 17 of 19 benchmarks.benchmarks·Seedream 5.0 Flash vs. Recraft: Image and Editing TestSeedream 5.0 Flash scored 8.0/10 for image generation and 9.2/10 for editing.benchmarks·H3 Max Relight, Turbo Extend and Pixelcut: Video Editing Testbenchmarks·Krea 3 Review: 18 Image Generation and Editing TestsI tested Krea 3 with 10 text-to-image prompts and 8 edits.benchmarks·SeedVR2 7B fp8: Upscaling 480p Video to 1080pI tested SeedVR2 7B fp8 on a 20-second AI clip.benchmarks·MOSS Image Review: Polish Test, Pricing and LicenseMOSS Image scored 8.5/10 in our Polish-language test.
benchmarks·How to Test AI Defending Networks When EDR Is AutonomousA new framework examines how to evaluate autonomous defense agents configuring commercial EDR systems.
benchmarks·Relational Model Training: 305 Seconds on One GPUbenchmarks·AI Lip Sync Test: Seedance 2.5, Wan 3.0 and InfiniteTalk
benchmarks·FairStress: Ai2 Benchmark Finds Hidden Bias in LLM Decisionsbenchmarks·Best AI for SQL: ChatGPT, Gemini or Claude Code?I tested GPT-6.1 Sol on a tricky SQLite database. It got all three queries right.benchmarks·Gemini 4 Argon Prompts: 12 Tested Examplesbenchmarks·Gemini 4 Argon vs Gemini 3.1 Pro: What’s Changed?Gemini 4 Argon scores higher and was faster and cheaper in a three-prompt Polish test.
benchmarks·COPEX: How MCP-Based LLMs Can Be Attackedbenchmarks·Grok 4.7 and AI Agents: 187 Emails, No Train RefundSix AI agents tried to recover a £350 train fare after the claim deadline.benchmarks·Mercury Decide vs Celeris-1: Speed and Benchmarksbenchmarks·Gemini 4 Argon for Coding: Test Results and Benchmarksbenchmarks·basal-omni: Remek Kinas’s Model Tops JevBench’s Open ModelsRemek Kinas’s basal-omni-1.0-ultra scores 82.5 on JevBench, ranking first among open models.benchmarks·Liquid AI d1-ultra: Decision Index Leader and Test ResultsLiquid AI's d1-ultra leads the Artificial Analysis Decision Index with 65.2 points.benchmarks·Perplexity Decider v1.1 27B: Cost, Benchmarks and APIPerplexity Decider v1.1 27B scored 53.1 in Artificial Analysis and costs $16.50 per million tasks.benchmarks·Microsoft-Decision-1: Ranking, Pricing and Test ResultsMicrosoft-Decision-1 scores 53.8 in Artificial Analysis.benchmarks·Kandinsky 6.0 Video: Local AI Video with Audio on RTX 4090Kandinsky 6.0 generates five-second video with sound.benchmarks·YuE2 Music Generation Test: 4 Local Models ComparedWe tested YuE2, ACE-Step 1.5, MiniMax Music 3 and DiffRhythm 2 with Polish and English lyrics on an RTX 4090.
benchmarks·VEED Dubbing on fal.ai: Polish Voice Cloning TestI tested VEED Dubbing on fal.ai with Polish and English clips.benchmarks·Gemini 4 Carbon benchmarks: all 19 tests and where it losesGoogle reports Gemini 4 Carbon leads in 17 of 19 benchmarks.benchmarks·Seedream 5.0 Flash vs. Recraft: Image and Editing TestSeedream 5.0 Flash scored 8.0/10 for image generation and 9.2/10 for editing.benchmarks·H3 Max Relight, Turbo Extend and Pixelcut: Video Editing Testbenchmarks·Krea 3 Review: 18 Image Generation and Editing TestsI tested Krea 3 with 10 text-to-image prompts and 8 edits.benchmarks·SeedVR2 7B fp8: Upscaling 480p Video to 1080pI tested SeedVR2 7B fp8 on a 20-second AI clip.benchmarks·MOSS Image Review: Polish Test, Pricing and LicenseMOSS Image scored 8.5/10 in our Polish-language test.benchmarks·How to Test AI Defending Networks When EDR Is AutonomousA new framework examines how to evaluate autonomous defense agents configuring commercial EDR systems.benchmarks·Relational Model Training: 305 Seconds on One GPUbenchmarks·AI Lip Sync Test: Seedance 2.5, Wan 3.0 and InfiniteTalkbenchmarks·FairStress: Ai2 Benchmark Finds Hidden Bias in LLM Decisionsbenchmarks·Best AI for SQL: ChatGPT, Gemini or Claude Code?I tested GPT-6.1 Sol on a tricky SQLite database. It got all three queries right.benchmarks·Gemini 4 Argon Prompts: 12 Tested Examplesbenchmarks·Gemini 4 Argon vs Gemini 3.1 Pro: What’s Changed?Gemini 4 Argon scores higher and was faster and cheaper in a three-prompt Polish test.benchmarks·COPEX: How MCP-Based LLMs Can Be Attackedbenchmarks·Grok 4.7 and AI Agents: 187 Emails, No Train RefundSix AI agents tried to recover a £350 train fare after the claim deadline.benchmarks·Mercury Decide vs Celeris-1: Speed and Benchmarksbenchmarks·Gemini 4 Argon for Coding: Test Results and Benchmarksbenchmarks·basal-omni: Remek Kinas’s Model Tops JevBench’s Open ModelsRemek Kinas’s basal-omni-1.0-ultra scores 82.5 on JevBench, ranking first among open models.benchmarks·Liquid AI d1-ultra: Decision Index Leader and Test ResultsLiquid AI's d1-ultra leads the Artificial Analysis Decision Index with 65.2 points.benchmarks·Perplexity Decider v1.1 27B: Cost, Benchmarks and APIPerplexity Decider v1.1 27B scored 53.1 in Artificial Analysis and costs $16.50 per million tasks.benchmarks·Microsoft-Decision-1: Ranking, Pricing and Test ResultsMicrosoft-Decision-1 scores 53.8 in Artificial Analysis.benchmarks·Kandinsky 6.0 Video: Local AI Video with Audio on RTX 4090Kandinsky 6.0 generates five-second video with sound.benchmarks·YuE2 Music Generation Test: 4 Local Models ComparedWe tested YuE2, ACE-Step 1.5, MiniMax Music 3 and DiffRhythm 2 with Polish and English lyrics on an RTX 4090.