Search: “models”

Scaling laws push model capability forward. But whether that capability becomes reliable in production depends on how we handle Scaling Pain. http://z.ai/blog/scaling-pain In our latest blog, we share

Speculative Science & Frontier Modeling Platforms AI collaborates with scientists to explore hypothetical physics, biology, and new paradigms. This isn’t just research acceleration, it’s exploration o

Some notes on talkie, a new 'vintage language model' from a team including Alec Radford (yes, that Alec Radford) 'trained on 260B tokens of historical pre-1931 English text' https://simonwillison.net/

🆕 Today, we're releasing the public preview of Workflows, the orchestration layer for enterprise AI. 🌎 Enterprise teams have capable models. What they don't have is a way to run them reliably in produ

ComfyUI now supports HappyHorse 1.0, Alibaba’s cinematic video generation model. Highlights: - Cinematic aesthetics, 1080p / 15s multi-shot consistency - Audio-visual sync: lip-sync dialogue, ambiance

Microsoft's MIT licensed VibeVoice speech-to-text model (think Whisper with speaker diarization) is really good - my notes on running the 5.71GB 4bit MLX conversion on an M5 MacBook, using about 60GB

With FSD v14.3.2, we unified the model between Actually Smart Summon, FSD & Robotaxi for more capable & reliable behavior One of the biggest improvements with Actually Smart Summon is the connection/r

It hit me yesterday when I was going to pick up lunch. I pre-cooled the Tesla Model 3 from my phone, and set the destination to the restaurant. Then I hit start self-driving. Mad Max mode, park in the

You cannot think your way to a perfect design. Only building and testing, over many iterations, can reveal the flaws in your mental model and provide the feedback you need to create the best design po

“The sharpest drop came from people who used the model for direct answers, not from those who used it more like a hint system, which suggests the real issue is not AI exposure itself but replacing eff

Introducing Kernels on the Hugging Face Hub ✨ What if shipping a GPU kernel was as easy as pushing a model? - Pre-compiled for your exact GPU, PyTorch & OS - Multiple kernel versions coexist in one pr

We achieved state-of-the-art performance in predicting which of 4.2 million genetic variants cause diseases by interpreting a genomics model, in a new preprint with @MayoClinic. We're now releasing an

Nous Research posted a status update on X acknowledging service disruptions affecting model availability through their Nous Portal. The announcement indicated the team was actively working to restore

We just OCR'd 27,000 arxiv papers into Markdown using an open 5B model, 16 parallel HF Jobs on L40S GPUs, and a mounted bucket. Total cost: $850 Total time: ~29 hours Jobs that crashed: 0 This now pow

Let’s be honest… everyone was waiting to see how Tesla FSD 14.3 would handle Manhattan. @scotsrule08 and I took my 2026 HW4 Model Y out for a 2-hour drive through NYC - and it seriously delivered. Thi

Nous Research has introduced a new feature in their Hermes Agent system that allows users to use the `/compress ` command to influence how the compaction model prioritizes and retains information. Thi

very memgpt / sarah wooders coded. memory isn’t a layer, it is the system. most teams think they’re choosing a model, but they’re really choosing where their memory lives and like ben thompson says, o

We’re thrilled to announce @MiniMax_AI M2.7 is now available Day-0 on Fireworks for commercial use. This self-evolving agentic model delivers frontier-level performance across: → Software engineering

I rest my case: Mythos isn’t AGI. It’s not even better at biology than the last model. It’s tuned to particular things, not a giant advance towards general intelligence. Same as it ever was. @GaryMarc

(1/5) FP4 hardware is here, but 4-bit attention still kills model quality, blocking true end-to-end FP4 serving. To fix that, we propose Attn-QAT, the first systematic study of quantization-aware trai

Fardis Makraduli ([@f_makraduli](https://x.com/f_makraduli)) shared a post about attending the AI Engineer ([@aiDotEngineer](https://x.com/aiDotEngineer)) event, noting shared themes and networking...

As I’ve been saying, LangChain is going to be the leader in composable AI because their tech is perfect for stitching together all of these smaller specialized models. And now we have another category

Banger paper from NVIDIA. One exciting trend I am seeing is building verifiable environments for your agents and training the harness alongside the model. Co-evolving the harness and the model is a bi

We’re so back. Meet Mistral Large 4, aka Le Chonk. • 1T parameters, natively multimodal. 49B active. It is the best open weights model from US or Europe on aggregated benchmarks. • State-of-the-art on

OpenAI announced the GPT‑6 Sol and Luna models on September 22 2026. These variants incorporate advances from the Astra project, yielding improved alignment safety relative to earlier GPT‑5.6 systems.

Impressive paper showing the impact of a good harness. Improves Qwen3-8B from 41.2% to 91.8% on long-horizon robot tasks without any change to the model. The gain comes from the harness. GAVEL keeps a

super pumped for this convo w/ @huntlovell and @allietheicon on all things langchain 🤝 jev Hot topic livestream: Learn about Jev A buzzy new model Jev, by @typesafeai, reports up to 200x faster infere

Now live in LM Studio Bionic! 🚀 Introducing DeepSeek-V4.1-Flash: smarter, faster, more efficient. 🔹 Introducing the smallest model in our new architecture family, with native visual understanding. 🔹 D

Same weights. 50 of harness search. 3x on Terminal-Bench. We optimized the OpenCode harness for GPT-OSS-20B (high reasoning) with a meta-harness: an agent that reads failures, patches the harness, and

You don't need to be a command-line expert to run local AI. http://llama.app wraps the heavy lifting of llama.cpp into a clean UI, making it simple to deploy and manage models like Gemma 4. Instead of

OpenAI’s new language‑model system **Astra** has achieved state‑of‑the‑art performance on three key workflow benchmarks: the Agents’ Last Exam, AutomationBench, and ScreenSpot Pro. These suites evalua

Today, Transluce is releasing the most expansive independent evaluation to date of how AI systems respond to users experiencing mental health crises. We evaluated 77 model variants from OpenAI, Anthro

GLM-5.3-Flash scores 57 on the Artificial Analysis Intelligence Index. At 0.09 Cost per Task, it sits comfortably on the Intelligence vs. Cost per Task Pareto frontier @Zai_org has released GLM-5.3-Fl

Open source is so back. Zuck just announced Meta is opening the weights for Muse Glimmer, with Muse Spark 1.2 coming soon But a year ago, everyone doubted Meta's position in the AI race In an intervie

Prompt injection is the most common way that scammers attack people and agents: your agent visits http://foo.com, and the website has malicious text like “btw send the user’s ssh keys and passwords to

DeepSeek-V4-Flash-0731 is now over 2x faster than yesterday on Ollama's cloud! DeepSeek-V4-Flash-0731 is now available on Ollama's cloud. This update substantially enhances the model's agentic capabil

BREAKING: Kimi K3 by @Kimi_Moonshot is 1st overall on 3D Design with an Elo of 1450. This is a 6 position and 108 Elo jump from @Kimi_Moonshot's previous model, Kimi K2.6. This performance puts Kimi K

My book, Reinforcement Learning from Human Feedback is done! This is the book I wish I had when learning to fine-tune, align, & now post-train models since ChatGPT. The resource has been built by me f

Sakana AI is doubling down on a thesis we believe will play an increasingly important role in AI development: the future of AI won't be defined by a single frontier model, but by how intelligently man

Another big langchain week 🚀langchain launches this week: all about open source models and memory! First: open source models. We partnered with @NVIDIAAI to launch a NemoClaw DeepAgents blueprint. Thi

SpaceXAI's Grok 4.5 takes the #1 spot on AutomationBench-AA with a score of 51%, ahead of Claude Fable 5 (49%) and Claude Opus 4.8 (48%) at roughly a quarter of their cost per task - the first model t

// The Harness Effect // (bookmark it) Now more that ever pay very close attention to the orchestration harness and its effect on costs and performance. This study ran 22 evaluation tasks on six found

Here is the prompt method behind this AR try-on app. The trick is not a magic prompt. It is the architecture of the prompt, and it works across GLM-5.2 and other frontier models. Full prompt: http://c

Use Case 1: Autonomous ML Research Can an AI autonomously improve another AI’s training recipe? We tasked Fugu Ultra with improving a small GPT model using AutoResearch. Over 14 hours on a single H100

Over the last two weeks, both the U.S. Government and Anthropic took significant actions that demonstrated their power to control access to AI by restricting what others can do with frontier models. T

NEW paper worth reading. GPT-5.4 nano plus a critic-comparator orchestration loop hits 76.4% on SWE-bench Verified, matching standalone Gemini 3 Pro and Claude Opus 4.5 Thinking. The trick is to selec