Category
Open Source AI
Local LLMs, open models, free AI infrastructure
// 10 articles filed
// FIG. CAT-OPEN
PristrenPristrenLocal LLMs, open models, free AI infrastructure
// 10 articles filed
Built by Pristren
Reading about AI tools? Run your team on Zlyqor — chat, meetings, projects, and time tracking in one workspace.
MPT-7B introduced ALiBi positional encoding for length generalization and shipped with an Apache 2.0 license, making it one of the first truly commercial-ready open LLMs.
Mahmudul Haque Qudrati
CEO & ML Engineer
Gemma 2 27B beats Llama 3 70B on MMLU (75.2% vs 73.1%) using knowledge distillation from Gemini and a novel sliding window attention design.
Mahmudul Haque Qudrati
CEO & ML Engineer
Llama 3.2 introduces vision capability to the Llama family with 11B and 90B vision models, plus 1B and 3B text-only variants for on-device deployment.
Mahmudul Haque Qudrati
CEO & ML Engineer
OLMo 2 is the only major LLM where you can reproduce the entire training run: weights, 3T-token Dolma dataset, training code, and evaluation suite are all public.
Mahmudul Haque Qudrati
CEO & ML Engineer
LLaVA 1.6 (LLaVA-Next) improves on its predecessor with dynamic high-resolution processing and 4x more instruction tuning data, achieving MMBench scores competitive with GPT-4V on several benchmarks.
Mahmudul Haque Qudrati
CEO & ML Engineer
Llama 3.3 70B closes most of the gap with the 405B model through better instruction following data and RLHF improvements - delivering 405B-class performance at a fraction of the serving cost.
Mahmudul Haque Qudrati
CEO & ML Engineer
Llama 3.1 405B achieves 88.6% on MMLU and matches GPT-4 on multiple benchmarks, with a commercial license for up to 700M MAU. Here's how to run it.
Mahmudul Haque Qudrati
CEO & ML Engineer
Nous Research's OpenHermes 2.5 demonstrates that one million carefully curated synthetic conversations can produce an instruction model that rivals much larger open weights.
Mahmudul Haque Qudrati
CEO & ML Engineer
HuggingFace H4 aligned a 7B model to beat Llama 2 70B Chat using only synthetic GPT-4 data and DPO - no reinforcement learning required.
Mahmudul Haque Qudrati
CEO & ML Engineer
Mistral 7B Instruct v0.3 delivers 32K context, function calling, and inference efficiency that rivals much larger models - here is how to deploy it.
Mahmudul Haque Qudrati
CEO & ML Engineer
Deep dives into ML algorithms, models, and applications
AI trends, techniques, and real-world implementations
How LLMs work, honest comparisons, and production usage
Every technique that works — with real examples
Claude Code, Cursor, Copilot, open-source tools reviewed honestly
Fewer tokens, cheaper APIs, local alternatives with real numbers
Benchmarks explained, evaluation frameworks, model testing
LLM SEO, AI SEO, Google AI Overviews, developer marketing
iOS, Android, and cross-platform mobile app development
Modern web technologies, frameworks, and best practices
Data analysis, visualization, and engineering insights
Autonomous agents, LLM applications, and intelligent systems