Lynn GitHub 镜像仓 · Qwen3.8-27B Coder390 EfficientThink 本地 GGUF（llama.cpp Q2/Q3，内置 Q4 MTP）· 主仓 https://github.com/MerkyorLynn/Lynn · 下载 https://download.merkyorlynn.com/download.html

electron benchmark typescript regression-testing huggingface ai-agent realtime-voice modelscope desktop-agent openai-compatible cli-agent agent-runtime coding-benchmark agent-eval agent-regression

• Updated Nov 12, 2026
• TypeScript

Pokedex benchmark for LLMs

pokedex cloudflare pokeapi ai-evaluation llm-benchmark coding-benchmark

• Updated Sep 13, 2026
• JavaScript

Benchmark of local LLMs (Qwen 3.6, Gemma 4) vs ChatGPT and Gemini on coding tasks. Apple M5, 32GB. Methodology, raw outputs, judge prompts, scores, and charts.

benchmark gemma mlx apple-silicon ai-evaluation llm local-llm ollama qwen coding-benchmark

• Updated Jun 2, 2026
• Jupyter Notebook

Reproducible frontier AI coding benchmarks with exact prompts, inspectable project snapshots, live demos and verifiable provenance.

benchmarking web-development reproducibility model-comparison ai-research ai-coding llm-benchmark ai-comparison coding-benchmark visual-benchmark

• Updated Nov 8, 2026
• JavaScript

Saotri Bench — coding benchmark for evaluating LLM agents on multi-phase programming tasks with hidden requirements.

python benchmark machine-learning code-generation evaluation-framework ai-agents llm llm-evaluation agent-evaluation coding-benchmark

• Updated Feb 21, 2026
• Python

Studiobook — SHIP's open benchmark fixture: a studio class-booking SaaS (Next.js) with seeded SDLC benchmark tasks

coding-agent coding-benchmark swe-in-a-team

• Updated Oct 31, 2026
• TypeScript

面向 AI IDE / 编码 Agent 的编程能力基准测试框架，内置防背题三防线：公私用例分离、泛化差距检测、参数化生成

python benchmark ai sandbox evaluation code-generation ai-agents anti-cheating llm ai-ide llm-evaluation coding-benchmark

• Updated Jun 6, 2026
• Python

31 local and hosted LLMs tested with one difficult falling-block browser game prom31 local and hosted LLMs tested with one difficult falling-block browser game prompt. Runnable outputs and hands-on findings.pt. Runnable outputs and hands-on findings.

javascript css html benchmark ai-evaluation llm local-llm coding-benchmark

• Updated Aug 22, 2026
• JavaScript

What extra reasoning buys in code and tests: 216 Sol coding attempts, MVCC defect detection, DAG execution, effort premiums and API-equivalent USD, with a reproducible stand and English report.

python reproducibility codex llm-evaluation coding-benchmark

• Updated Oct 5, 2026
• Python

Contamination-proof LLM benchmark: fresh, verified coding tasks every run, impossible to benchmaxx

benchmark data-contamination llm evals openrouter llm-evaluation coding-benchmark benchmaxxing

• Updated Oct 4, 2026
• TypeScript

Raw logs of Claude Code running on local Qwen3.5-27B (llama.cpp). Builds a Python todo app with 50 tests. Real-world performance data: 30 min, cache thrashing, 38 t/s generation.

reproducible-research wsl performance-analysis unit-tests python-todo edge-ai ai-tools ai-agent llm llama-cpp local-llm local-ai qwen ai-productivity opensource-ai ai-benchmark claude-code agentic-coding privacy-focused-ai coding-benchmark

• Updated Apr 14, 2026
• Python

Benchmark and tune every locally installed Ollama model on your own GPUs. Finds the best model per GPU setup plus its optimal temperature and context, with a no-spill VRAM guard, coding + agentic test suites, and a single-file HTML report. 100% local: no cloud, no API keys, no model judges.

python windows vulkan self-hosted quantization multi-gpu model-evaluation amd-gpu inference-optimization gpu-benchmark llama-cpp local-llm local-ai ollama gguf agentic-ai ai-benchmark llm-benchmark llm-tuning coding-benchmark

• Updated Sep 17, 2026
• Python