Tag

#open-source-ai

9 posts

NewsAugust 14, 202617 min
NVIDIA NeMo Switchyard: Agent Model Routing 2026
NVIDIA's NeMo Switchyard routes each step of an AI agent to a different model. LangChain's benchmark found 74% lower cost — and a six-point accuracy drop.
AI & Machine LearningApril 19, 202611 min
GLM-5.1: The Open-Source Model That Beat GPT-5.4
GLM-5.1 (April 2026): Z.ai's 754B open-weight model scored 58.4% on SWE-bench Pro, beating GPT-5.4 and Claude Opus 4.6 on real coding benchmarks.
Security & PrivacyApril 12, 202611 min
AI Cybersecurity's Jagged Frontier: Small Models vs Mythos
AISLE tested 25+ AI models against Mythos's showcase vulnerabilities. A 3.6B model found the same FreeBSD flaw. Here is what the jagged frontier means.
AI & Machine LearningApril 5, 202611 min
AirLLM Tested: Run a 70B LLM on a 4GB GPU — Does It Work?
Run a 70B LLM on a 4GB GPU? AirLLM uses layer-wise inference, no quantization. Benchmarks, latency tradeoffs, and how it compares to Ollama + llama.cpp.
AI & Machine LearningApril 4, 202613 min
Gemma 4 Review: Apache-2.0 Open AI, 89.2% on AIME 2026
Google DeepMind's Gemma 4 family: 31B Dense scores 89.2% on AIME 2026, 26B MoE activates 3.8B/token, E2B fits in 1.5 GB. Apache 2.0, runs on your GPU.
AI & Machine LearningMarch 6, 202611 min
OpenCoder Review: Free Open-Source Code Model
OpenCoder review: the free-to-download code model in 1.5B and 8B variants. 83.5% HumanEval, 79.1% MBPP — deployable on your own hardware, though the model weights carry a custom license, not Apache 2.0.
LLM & IntegrationMarch 2, 202611 min
LM Studio 2026: Run Local LLMs With GPU Acceleration
LM Studio runs open-source LLMs locally on Windows, Mac (Apple Silicon), and Linux. Setup, GPU (CUDA/Metal/Vulkan/ROCm), model picks, and RAG in one guide.
Security & PrivacyOctober 1, 20259 min
AI SOC: Security Operations with AI Agents
AI SOC: how intelligent agents reshape the Security Operations Center. Alert triage, automated response, and the tooling ending the alert-fatigue era.
LLM & IntegrationSeptember 28, 20259 min
Moonshot’s Kimi-K2: The Open-Source AI Model Beating Paid Giants
Moonshot Kimi-K2: the open-weight, trillion-parameter MoE model that outscored GPT-4.1 on SWE-bench Verified at release. What it is, and what it actually ships.

Every post in #open-source-ai (9)