Search: “models”

Hugging Face and AMD have collaborated to quantize the Kimi-K2.6 model to MXFP4 format, enabling efficient inference on AMD hardware. MXFP4 is a low-precision floating-point quantization technique tha

Gary Marcus critiques ChatGPT's lack of embodied understanding and spatial reasoning, arguing that the language model struggles with physical concepts that humans intuitively grasp through bodily expe

Could a sequence of poetic gibberish manipulate the world’s most advanced AI? Recent tests on GPT-5 by @ChristophHeilig show it consistently rates nonsensical 'word salad' as higher quality than clear

Excited to share some big news: I've joined @FireworksAI_HQ President. I've spent my career at the intersection of great teams, disruptive technology, and massive market opportunity, from architecting

Guys, I am absolutely astounded. The Qwen 3.6 27b is like a jump to Qwen 4 from Qwen 27B 3.5. I just did a full suite of front end design tests and agentic benchmarks, made entirely by it. VERDICT: Th

How far are we from agents that can self-generate world knowledge? The work proposes an outcome-based reward that measures how much an agent's self-generated world knowledge actually improves its task

I predicted in January that 'harness' would be the AI buzzword of H12026 - seems accurate so far. Everyone is talking about harnesses but I'm not confident that the majority of people (including me!)

LM Performance：With only 27B parameters, Qwen3.6-27B outperforms the Qwen3.5-397B-A17B (397B total / 17B active, ~15x larger!) on every major coding benchmark — including SWE-bench Verified (77.2 vs.

Q1 2026 Shareholder Update https://ir.tesla.com/#quarterly-disclosure We continued to make meaningful progress on the build out of the infrastructure & AI software that underpins our Robotaxi & future

Thank you for flagging this, Jeff. This was a mistake: we are not deprecating text-embedding-3-small. We’re looking into where this came from now, and we’ll also email users to clarify. Sorry for the

Claude Opus 4.7 with adaptive thinking via the API... am I missing something or is it not possible any more to force it to think? (Prompt hacks like 'think step by step' don't count here, I mean the e

fun off-the-cuff harness + agent engineering sesh with @dexhorthy @vaibcode @GeoffreyHuntley - a lot of good harness/agent design we’re doing today is still good context engineering, remember tool cal

memory ended not being valuable in ChatGPT possible this exact implementation isn't valuable either but some memory implementation will be valuable, will create a bunch of lock in. and everyone is rus

ParseBench is the first benchmark to include VLM chart understanding 📊📈📉 over enterprise documents. 🟠 Existing benchmarks (ChartQA, ChartXiv) test over charts specifically and not the chart's inclusio

Classic study gave 146 economist teams the same dataset & got wildly different answers New paper reruns it with agentic AI. Claude Code & Codex land near the human median, but with far tighter dispers

Kimi K2.6 just entered the chat. > beats gpt 5.4 and opus 4.6 on coding. > open source. open weight. > 3x-5x cheaper. > really good for people running agents. > insane at running long tasks. > we're t

Kimi-K2.6 is a new AI model released through Qoder that achieves state-of-the-art performance in coding tasks and supports long-horizon execution and agent swarms, available at a reduced cost of 0.3x

A user reports that Opus 4.7 performs better in Claude Code when executed with a minimal system prompt (using just a period) rather than the full default system prompt, suggesting that reducing system

This post discusses advanced rendering techniques for realistic visual effects, including sheer fabric simulation with light transmission properties, cloth physics responsive to wind forces, depth-of-

ComfyUI's HY 3D 3.0 nodes enable rapid 3D asset generation from single or multi-view images, producing high-fidelity 3D models suitable for applications including product visualization and game develo

We have been named one of the 40 Most Innovative AI-Native Prosumer Companies by @notablecap. The Prosumer AI 40 recognizes companies building the tools that blur the line between professional and con

🔴 BREAKING Speculative decoding just got real for local LLMs. llama.cpp merged speculative checkpointing (PR #19493). Expect 0-50% speedups on coding tasks with `--spec-type ngram-mod`. This means: •

Just got access to Grok 4.3 and the first thing I did was throw the hardest task I usually run on claude to test it gave it a complex research prompt on deepseek: >thought for 13 mins and 17 seconds >

OMG. Let’s get one thing straight. Claude doesn’t get anxious. It mimics people who get anxious. Those two things are NOT the same. My head is shaking so much I need medical attention. anthropic's in-

This post compares AI assistants' responses to a request about driving to get gas, claiming that Grok 4.3 provides the requested information while ChatGPT, Claude, and Gemini decline or suggest altern

GDPval is one of the most important benchmarks of AI ability because it is based on human expertise. It compares expert human performance to AI performance using expert human judges who spend an avera

I prefer my design tool to be more closely integrated with where my agents work. I spent a few hours building my own design tool (inspired by Claude Design) inside my orchestrator. I can use this with

I think Artificial Analysis does a good job overall and provides transparency in benchmarking, but GDPval-AA is not a good benchmark and needs to stop being reported. It is Gemini 3.1 judging other mo

NEW paper from Apple. Interesting idea: 'Attention to Mamba'. The paper introduces a two-stage recipe for cross-architecture distillation from Transformers into Mamba. Naive distillation collapses tea

30K+ likes in the first hour. 🤯 That is crazy! Design is unsolved with agents. But lots of impactful work generated by agents is around design. Claude Design is Anthropic's way of saying that they are

Hermes creates skills from experience, improves them during use, nudges itself to persist knowledge, searches its own past conversations, and builds a deepening model of who you are across sessions. D

in 1996 mitzenmacher showed that sampling two backends and picking the better one drops max load exponentially vs. random selection one extra comparison. that's the whole trick we built pinecone assis

Agent evals are drifting away from production reality. Most benchmarks use clean tasks, well-specified requirements, deterministic metrics, and retrospective curation. Production work is messier, with

Anthropic says Opus 4.7 hits 80.6% on Document Reasoning — up from 57.1%. But 'reasoning about documents' ≠ 'parsing documents for agents.' We ran it on ParseBench. → Charts: 13.5% → 55.8% (+42.3) — h

Claude Opus 4.7 is now available in Windsurf 2.0! Anthropic has clearly optimized Claude Opus 4.7 for sustained reasoning over long runs. Agents stay on track longer without intervention, so engineers

Claude Code users should ensure they are using Opus 4.7 with the latest updates to experience performance improvements or feature enhancements. The post suggests that users not observing expected incr

Nous Research shared an update or announcement on their official X (Twitter) account, likely relating to their ongoing work in AI model development, fine-tuning, or research releases. Nous Research is

in the grand narrative of Meta x AI, we saw the flop (Llama 4 hurhurhur), and now we’re seeing the turn: - *more* hiring since the soup wars of 2025 - Zuck literally moved in with Alexandr and Nat and

This post celebrates Claude Opus 4.7's capabilities within Claude Code, expressing enthusiasm about its performance and describing it as exceptionally powerful. The entry references an Anthropic annou

🎬 Ollama Gemma Day Recap: SGLang at the Ollama Gemma 4 Party in Palo Alto 🍾 Last night, @ollama hosted a packed Gemma Day at the Palo Alto office alongside the @GoogleDeepMind Gemma team. SGLang was i

Opus 4.7 is in Claude Code today. It's more agentic, more precise, and a lot better at long-running work. It carries context across sessions and handles ambiguity much better. Introducing Claude Opus

Shocking result on my pelican benchmark this morning, I got a better pelican from a 21GB local Qwen3.6-35B-A3B running on my laptop than I did from the new Opus 4.7! Qwen on the left, Opus on the righ

There's a broadly held misconception in AI that methods that scale well are simple methods -- even, that simple methods usually scale. This is completely wrong. Pretty much none of the truly simple me

Today, Clearwing has replicated Glasswing's discovery of the @FFmpeg vulnerability using Codex-5.4 as the backend. I am going to attempt the same today, using @Alibaba_Qwen 3.6 on @ollama running on m

We’re back 🔥. Thrilled to be named once again to the @Forbes AI 50. The AI Native Cloud, built for the full AI lifecycle: fast inference, open models, fine-tuning at scale. https://www.forbes.com/list

What you need to know about Opus 4.7 * Takes instructions literally * Better vision means improved computer use and producing slides and other visual artifacts * Optimized for large-scale real-world a