Videos
Find videos about LLM models, hardware, runtimes, benchmarks and practical AI workflows.
Muse Glimmer and Qwen 3.8 27B implement two new features in SuperAsteroids, scored by code review with merging or deleting branches based on results.
Tests experimental Apple Neural Engine Prefill for Qwen3.8-27B on Apple Silicon via oMLX, comparing performance with ANE enabled versus disabled during the prefill stage.
Tests Ox Alpha via OpenRouter using motorcycle simulation, watch website creation, 3D CAD modeling, creative frontend design, ship combat simulation, and multimodal coding tasks.
Thorough testing of DeepSeek-V4-Flash-Vision-Exp.
Tests Ornith-1.5-35B-A3B-GGUF performance, memory, agency, OpenAI Human Eval, Kanban, Sand Physics, Dungeon Crawler, Blender and Godot.
Covers Ox Alpha stealth model, GLM-5.3 Flash, Tencent Hunyuan HY4, delayed GPT-6 Astra, held Fable 5.1, Codex limits, Anthropic retention changes, and OpenAI Mona Lisa image model.
Evaluates optimal KV cache quantization using Bob-bench on Mac Studio and 4090 hardware across varying context lengths, referencing a Meta paper.
Installs DeepSeek Harness on Linux, wiring it to Ollama and Unsloth servers running Qwen 3.8-27B at 4-bit locally, then builds a to-do list app in 30 minutes at 16 tokens per second.
Installs and tests Escha-W2, a 2-bit quantized build of Qwen3.8-27B.
Forces BearQ's Explorer Agent to instantly detect new EAApp search features, generate tests via permutations, and update the Application Model using QA Lead and Tester agents.
Evaluates Ornith-1.5-397B-MLX-Q9 performance using Inferencer App on M3 Ultra 512 GiB against claims of surpassing GLM 5.2, Opus and Qwen 3.8 MAX.
Tests dots3-note prev, identified as the first open-weight model in the dots3 family.
Tests Qwen 3.8 27B Q4_K_M reliability via clock tasks, Pomodoro timers, and building a Feed Aggregator in Nim using Ollama and Pi agent.
Tests Qwen3.8-27B local inference on Apple M5 Max, measuring TPS across reasoning levels and evaluating DFlash2 speculative decoding acceleration combined with varying thinking modes.
Locally installs and tests the Ornith-1.5-9B model using resources from huggingface.co/ornith-ai/Ornith-1.5-9B.
Demonstrates building ChatPromptTemplate with runtime placeholders, handling missing variables and JSON curly brace traps via double escaping, partial fills with .partial(), and integrating chat histo
Tests using DeepSeek Harness to call Claude Code and Pi as sub-agents, covering plugin architecture, modes, and custom plugin creation.
Locally installs and tests Ornith-1.5, a model built through end-to-end self-improvement.
Thorough testing of GLM 5.3 covers coding capabilities, security aspects, and creative coding applications using z.ai resources.
Explains QWEN3.8 performance causes, benchmarks it live on M5 Max, and demonstrates replication steps for identical results on local setups.
Tests how Qwen 3.8 27B reasoning levels impact output quality and token usage across tasks including stopwatch, typing speed, sand physics, dungeon crawler, notes app, blender windmill, and charts.
Tests Qwen 3.8 27B against Opus 4.6 on a live game development task using an existing codebase hosted at brrnout.com.
Evaluates Qwen 3.8-27B reasoning_effort levels via Tetris clone, volcano sim, spreadsheet, and drawing tests using identical official thinking samplers on a 17.9 GB 4-bit file.
Local installation and testing of Qwen3.8-27B using Dflash2, featuring live local benchmarking performance evaluations.
Demonstrates Qwen3.8-27B running unsupervised local coding tasks for 10 hours, including building a Pi/goal extension via Grok Build and optimizing SGLang-Omni to serve MiniMax-Music3.
Installs and tests hermes agent loop functionality using Qwen3.8 27B to demonstrate recurring execution of prompts or slash commands.
Evaluates Grok 4.6 and Grok Bot against Fable 5, GPT 5.6, and Hermes Agent via BridgeMind 1 tasks, fal.ai plugins, Remotion teasers, and crash fixes using speed and cost metrics.
Explores TalkWithMe v4 features including voice cloning, persona editor, echo chamber and true group chat using dots.tts or Qwen3-TTS.
Demonstrates SystemMessage, HumanMessage, and AIMessage mapping, LangSmith tracing setup, pirate/support/rhyme prompt tests, token cost measurement, and prompt injection against a no-refund agent usin
Demonstrates generating test cases from requirements, authoring and executing scripts, debugging failures, and running on 30,000+ real devices via Playwright, Selenium, and Cypress within the IDE.
Installs and tests Qwen3.8-27B-Ridge-3.7bpw using GGUF format from empero-ai repository.
Compares Qwen3.8 27B and Qwen3.6 27B using Regina on Minesweeper, VMs, py2c calc, encryption and memcached, analyzing reasoning tokens, degeneracy and failures.
Evaluates Qwen 3.8-27B quantizations (BF16, Q8, Q4 Unsloth, Q4_K_M Ollama, Q2) via one-shot exams measuring speed and quality against a BF16 baseline.
Installs free open source LTX 2.5 in ComfyUI, demonstrating text to video, image to video, first last frame generation, using loras, low VRAM settings, Minimax H3 vs LTX 2.5 comparison, Minimax plus L
Tests Qwen 3.8 using Inferencer App v2.3.3 on M3 Ultra 512 GiB across 2 million tokens, varying temperature, quantization levels, and thinking modes.
Installs and tests Antares-1B for vulnerability localization in real-world codebases.
Evaluates Qwen 3.8 27B using an eight-stage development plan on improved hardware achieving 70–130 tokens per second, with code review included.
Evaluates Qwen3.8-27B inference via llama.cpp RPC across RTX 3060 and Apple Silicon Mac, comparing throughput metrics for single-card versus distributed layer loading.
Tests Gemini 3.7 Flash on browser workflows, 3D scenes, C++ games, CAD models, hardware mods, FPS dev, frontend design, and the Street Yeet game.
Demonstrates making a brownfield codebase AI native across PM, developer, and QA roles using the AI Layer Starter Pack, global rules, and automated PR reviews in CI.
Tests DeepSeek V4 Pro (0813) via WoAI Bench on frontend dev, agentic coding, 3D/Three.js, and one-shot gen, comparing cost, speed, and quality against Gemini 3.7 Flash, Grok 4.6, Kimi K3, GLM 5.2, Qwe
Installs and tests Qwen3.8-9B, described as a full-parameter distillation of Qwen3.8 2.4T A95B.
Qwen 3.8-27B versus Qwen 3.6-35B-A3B at BF16 on a custom exam featuring Tetris, volcano physics, spreadsheets, and SVG tasks, plus Claude Opus 4.6 comparison.
Reviews GLM 5.3 capabilities through demonstrations involving Z Code, Windows replica creation, Blender V8 engine usage, 3D fighting games, music composition, financial presentations, deep research, c
Evaluates DeepSeek V4 Pro and V4 Flash via BridgeMind horror game, Minecraft clone, Next.js dashboard, and BridgeMind 1 bug against GPT 5.6, Fable 5, and Grok 4.6 using DeepSeek Harness.
An app based on local AI benchmarks assists users in determining suitable hardware configurations for running specific machine learning models locally.
Installs and tests LTX-2.5, described as an open world model, within ComfyUI, covering setup fixes and initial generations using resources from huggingface.co/Lightricks/LTX-2.5.
Configures Claude Code to use Nemotron 3.5 Lightning via Ollama on an RTX 5090 through .claude/settings.local.json, verified by generating and modifying a Python palindrome function.
Tests GLM 5.3 on browser workflows, C++ game development, 3D CAD modeling, frontend design, FPS generation, and a difficult hardware test.
Demonstrates installing Unsloth Studio via terminal on AMD ROCm hardware, loading the full 55 GB Qwen 3.8-27B model consuming 83 GB RAM, and testing local chat, vision, and deep research capabilities.