Muse Spark Open: Meta's 816B Open Model Runs Like a 42B One
Muse Spark Open is Meta's open-weight agentic coding model: 816B parameters, 42B active, 1M context. What it takes to run it, and who should try it first.
Nemotron 3.5 Super VL: Nvidia's Open Model for Always-On Agents
Nemotron 3.5 Super VL is Nvidia's open model for always-on enterprise agents: 120B parameters, 12B active, open data. What it means for self-hosted agents.
LFM2.5-24B-A2B: An On-Device Agent Model With 2B Active Params
LFM2.5-24B-A2B runs on-device agents with only 2B active parameters, leading on tool use but trailing on coding. What the benchmarks show and where it fits.
Reflection's Beam: The Open-Weight Benchmark That Closed a Gap
Beam matches GLM-5.2 at 3–4x less compute, ships Apache 2.0, and emerged web browsing unprogrammed. What open-model practitioners need to know.
Alibaba's Qwen-Drive Can Tell You Why It Braked — The Explanation Just Might Not Be True
Qwen-Drive 1.0 is an open-source autonomous driving model that generates natural language explanations for its decisions. Its own researchers found those explanations don't reliably match the actual decisions.
Nvidia's $12.9B Hugging Face Acquisition: Open Source Becomes the Strategic Choke Point
Nvidia is paying 80x revenue to acquire Hugging Face as OpenAI and Anthropic build custom chips — making open-source developer mindshare the most valuable asset left in AI infrastructure.
Mojo Is Now Open Source — What It Means for GPU Programming
Mojo's compiler and toolchain went Apache 2.0 on August 18, fulfilling a promise made since May 2023 and opening GPU-optimized Python-syntax programming to community development.
DeepSeek Open-Sources Its Agent Harness — and Raises API Prices on the Same Day
DeepSeek shipped a better V4 Pro, released its agent framework under MIT, and raised API prices by up to 4× — all at once. Here's what each move means for teams building on open-weight models.
Nvidia's Nemotron 4 Aims for One Trillion Parameters — Chinese Labs Already Passed That
Nvidia is developing a 1-trillion-parameter open-weight model targeting a benchmark that Kimi K3 (2.8T) and DeepSeek V4 Pro (1.6T) have already surpassed. The race is interesting for a different reason.
Meta's Muse Glimmer Bets on Local Agentic AI Over the Cloud
Meta's new open-weight Muse Glimmer model runs agentic tasks on a single consumer GPU without cloud dependencies. Zuckerberg frames it as a direct challenge to AI power concentration.
FLUX 3 Video: Black Forest Labs Claims Top Elo Score at $0.06 Per Second
FLUX 3 Video is now generally available via API with a 1,135 Elo score, 14-language lip sync, native audio, and full HD output — all included from $0.06 per second.
Kimi K3's Weights Are Open. Its Commercial License Just Got Tighter
Moonshot's Kimi K3 ships at frontier pricing on OpenRouter, but the real news is a license that closes the free-ride window at $20M in trailing revenue.
Cisco's Antares: 150x More Bugs Per Dollar Than GPT-5.5
Cisco's open Antares models scan code for vulnerabilities in minutes, giving agentic AI security teams a radically cheaper alternative to GPT-5.5.
Nativ Turns Any Apple Silicon Mac Into a Local AI Server
Prince Canuma's open-source Nativ app finds AI models already cached on your Mac and serves them locally for chat and coding agents — no cloud, no account.
Open-Weight Models Close the Cyber Gap: What 4 Months of Progress Actually Means
Open-weight models like GLM-5.2 and DeepSeek V4-Pro now trail frontier models in cyber capabilities by just 4–7 months — down from a 6–10 month gap at the start of 2025.
Bonsai 27B: A 5.9GB LLM That Keeps 95% of Its Reasoning on Your iPhone
PrismML's Bonsai 27B compresses Alibaba's Qwen3.6-27B to 1–2 bits via ternary quantization — 5.9GB for the standard model, 3.9GB for the iPhone variant. Math and coding benchmarks are virtually unaffected.
Chinese AI Models Now Top 30% of OpenRouter Traffic as Cost Gap Hits 90%
Chinese open-source models now claim over 30% of OpenRouter's weekly API traffic, running 60–90% cheaper than US counterparts. Here's what the pricing shift means for teams choosing their AI stack.
Ornith-1.0: Open-Weights Self-Scaffolding LLMs That Handle Agentic Coding Natively
DeepReinforce's Ornith-1.0 brings self-scaffolding to open-weights models from 9B to 397B — MIT licensed, built on Gemma 4 and Qwen 3.5, with state-of-the-art coding benchmark results.
GLM-5.2 vs. Opus 4.7: How China's Pricing Is Rewriting AI Cost Expectations
Snowflake's CEO ran a real-world coding benchmark where GLM-5.2 matched Opus 4.7 at 80% lower output token cost. What the numbers actually show — and what they miss.
Xiaomi's MiMo Code Benchmarks Past Claude Code on Terminal Bench — as Open Source
Xiaomi released MiMo Code on June 10 — an open-source agentic coding agent built on a 1T-parameter MoE model. Terminal-Bench 2.0: 86.7% vs Claude Code's 65.4%. Past 200 steps, it posts a 65%+ win rate in human A/B testing with 576 developers. The self-reported benchmark caveat applies.
Kimi K2.7 Code: Open-Weights That Undercuts Claude by 12×
Moonshot AI's newest coding model prices output at $4.00/M tokens versus Claude Fable 5's $48. Here's what that cost gap means for agentic pipeline economics.
agentgateway Joins AAIF: The MCP Router With 300+ Contributors Gets a Linux Foundation Home
agentgateway—the open-source gateway for MCP, agent-to-agent protocols, LLM inference, REST, and gRPC—joins the Agentic AI Foundation as its fourth hosted project, bringing enterprise governance and observability to agentic traffic.
Why Large Models Learn Rare Skills That Small Ones Forget
New research pinpoints the mechanism behind emergent LLM abilities: an update-and-forget pattern where larger models preserve rare-task signals that smaller ones lose when frequent tasks dominate training steps.
Ideogram 4.0: The First Open-Weight Image Model That Can Actually Render Text
Ideogram 4.0 releases as an open-weight model with native 2K resolution and top-ranked text rendering. It outperforms Midjourney v8 on DesignArena and enables fine-tuning on proprietary datasets.
This Open-Source Voice Model Listens Nonstop and Decides Every 0.4 Seconds Whether to Respond
Audio Interaction is a new open-source voice model that makes real-time decisions every 0.4 seconds whether to speak or stay silent — no waiting for you to finish, no push-to-talk, no submission step. Available now under Apache 2.0.
MiniMax M3: The Open-Weight Model That Runs Million-Token Contexts
MiniMax M3 is the first open-weight model to achieve 1M token context at proprietary-class performance — beating Opus 4.7 on autonomous web search while running 4x faster.
Nous Research Releases Hermes Desktop: Open-Source AI Agent for 300+ Models
Nous Research launched Hermes Desktop, an MIT-licensed AI agent for Windows, macOS, and Linux that runs across Telegram, Discord, Slack, email, and the terminal — with access to over 300 models via Nous Portal.
DeepSeek Makes the 75% Discount Permanent: What $0.87/M Output Tokens Changes
DeepSeek made its 75% discount permanent — output tokens at $0.87/M versus GPT-5.5's $30/M. The strategic case for 'good enough' model routing.
Alibaba Qwen3.7-Max: 35 Hours Autonomous, 10x Chip Speedup, Zero Documentation
Qwen3.7-Max ran 35 hours autonomous on a chip it had never seen, making 1,158 tool calls to reach 10x speedup. What this means for production agentic AI.
OPPO Open-Sources X-OmniClaw: The On-Device Android AI Agent
OPPO's X-OmniClaw keeps sensing on-device—cloud only receives planning queries. A reference architecture for private agentic AI on Android.
DeepSeek V4's 1M-Token Context Uses 10% of V3's KV Cache
DeepSeek V4 runs 1M-token context at 10% of its predecessor's KV cache. Gemma 4 does 128K on-device. Both target the same bottleneck: KV memory.
DeepSeek V4 Pro: Frontier Coding Performance at 7× Lower Cost Than Claude Opus
DeepSeek V4 Pro scores within 0.2 points of Claude Opus 4.6 on SWE-bench Verified at $3.48 per million output tokens vs. $25. The 7× price gap reshapes the agentic AI infrastructure cost model.
Gemma 4: Google's On-Device AI That Selects Its Own Tools
Google's Gemma 4 runs on Android and iOS with no data leaving the device — and autonomously activates agent skills without manual tool selection. Here's what the architecture means for mobile agentic AI.
GLM-5.1: The First Open-Weight Model to Lead SWE-Bench Pro
Z.AI's 754B MoE model tops SWE-Bench Pro above Claude Opus 4.6 and GPT-5.4, sustains 8-hour autonomous operation, and ships under MIT license. Here's what that combination changes.
Google AI Edge Gallery: Run Gemma 4 on iPhone With Built-In Tool Calling
Google's official app lets you run Gemma 4 models locally on iPhone, complete with working tool calling across eight widgets. It's the first official on-device AI app from a major model vendor.
Xiaomi MiMo V2: Three AI Models Built for Agents, Robots, and Voice
Before its official launch, Xiaomi's flagship model ran anonymously on OpenRouter, topping daily rankings. Now MiMo V2 Pro, Omni, and TTS reframe what Chinese AI infrastructure means for agentic workloads.
TADA: The Open-Source Speech Model With Zero Hallucinations
Hume AI's TADA runs 5x faster than comparable speech models, hallucinates zero words across 1,000+ tests, and fits on a smartphone. The key is a 1:1 audio-to-token mapping.
Nvidia's $26B Open-Source Bet That Could Reshape Agentic AI
Nvidia is spending $26 billion on open-weight AI models. Here's what that shift means for agentic AI infrastructure and the developers who depend on it.
Helios Makes Minute-Long AI Video a Real-Time Problem
ByteDance's open-weight Helios model generates minute-long video at 19.53 FPS on a single H100 — faster than any competing 14B model and cheap enough to deploy seriously.
The Qwen Exit: Open-Source AI Strategy and the Risk of Leadership Vacuum
Analysis of the mass departure from Alibaba's Qwen team following the release of Qwen 3.5 and what it means for the future of open-weight models.
Qwen 3.5: The New Open-Weight Benchmark for Function-Calling Agents
Analysis of the Qwen 3.5 release and its implications for function-calling efficiency and high-density reasoning in the Agentic Stack.