A user demonstrates running the Qwen 3.8 27b model with llama-server to generate an HTML simulation of an aquarium panel failure, including detailed token usage and model parameters.
After: https://www.reddit.com/r/LocalLLaMA/comments/1vckcue/ds4_flash_0731_acquarium_panel_failure_q3_k_xl/ start C:\llm\llamaM6\build\bin\llama-server.exe --model "F:\models\Qwen3.8-27B-UD-Q8_K_XL.gguf" --temp 1.0 --top-p 0.95 --top-k 20 --min-p 0.0 --presence-penalty 0.0 --reasoning-preserve --repeat-penalty 1.0 --ctx-size 248320 --no-warmup --no-mmap --fit on --parallel 1 --cont-batching --reasoning on --n-cpu-moe 0 -lv 4 --device CUDA0 --mmproj "F:\models\mmproj-BF16.gguf" --spec-type draft-mtp --spec-draft-n-max 3 --chat-template-file f:\models\chat_template.jinja RTX6000 96gb ___ Prompt: Create a large glass aquarium whose side panel develops a visible crack and then bursts. The simulation must include: Water escaping through the opening with flow strength based on water depth and decreasing as the tank drains A curved water jet affected by gravity A spreading puddle that collides with the room boundaries Fish, rocks, plants, and a floating toy reacting differently according to density, buoyancy, drag, and current Objects transitioning correctly from underwater motion to airborne motion and then to floor collisions Fish attempting to swim against the current before being swept through the breach Glass fragments with angular velocity, collisions, and water resistance A visible waterline that lowers continuously rather than disappearing all at once Let the user drag the crack vertically before triggering the failure. A lower crack should initially produce a stronger jet than a higher crack. Give me 1 html file _________________________________________________ Total tokens: cross all turns Prompt tokens evaluated 84.250 tok 1012 reused from KV cache Tokens generated 116.080 tok This turn · KV cache Prompt 84.246 tok 83.234 fresh + 1012 cached Generated 32.877 tok KV cache total 117.123 tok Avg speed 74.2t/s 2 turns. Second turn prompt: Bug: the fish and the duck do not come out of the aquarium.
Tested a prompt on the Qwen 3.8 27B model using GitHub Copilot's agent mode to simulate a detailed aquarium burst scenario, demonstrating iterative code generation without user input.
A user describes an unexpected behavior where the quantized Qwen 3.8 27B model opened a browser to test code autonomously during development, highlighting emergent capabilities.
This article showcases the output of the Qwen 3.8 FP8 model in generating a self-contained HTML file for an animated lava lamp, highlighting its impressive code generation performance.
The article showcases the Qwen 3.8 27b AI model's ability to generate a sophisticated JavaScript and WebGL project for realistic real-time ocean rendering in the browser.
A user praises the Qwen3.8-27b model (q3_k_xl quant) for reliably wiring up OpenWebUI, spawning services, fixing bugs, and debugging a Cloudflare tunnel, highlighting Unsloth's low-bit quantization as enabling surprisingly capable local utility use.