Search: “models”

Simon Willison requested that Hugging Face add a repository size sorting option to help users find and filter models by storage requirements, specifically mentioning interest in locating DeepSeek quan

GPT-5.5 is a giant leap forward for handling ambiguity compared to previous GPT models. As Windsurf 2.0 focuses more on parallel agents, this model is key for long-horizon tasks — it excels at underst

Love this work from Aksel and the post-training team at Hugging Face! Turns out the HF ecosystem (papers, datasets, models all accessible through CLI, skills and md files) is perfect for running SOTA

I gave two MoE models the same vibe coding challenge Qwen3.6 35B A3B (31.8GB) vs Gemma4 26B A4B (23.3GB) Stack: > Unsloth Q6_K_XL > llama.cpp > Model-card recommended sampling for each 4 prompts, side

Fireworks AI published a compilation of insights on achieving training-inference parity in Mixture of Experts (MoE) models, addressing fundamental challenges in aligning model behavior between trainin

We've tested new OSS models the moment they're released for a while at Lindy. Inference is our #1 cost by a lot (more than payroll) — cutting it by 2-5x would be transformative. Last year, OSS models

OpenAI should probably bite the bullet and just name their next set of models something more human sounding. Everyone anthropomorphizes their AIs anyway, and 'Claude' is an easier name to refer to tha

The founder of a $4B inference company says that if you're building agents, foundational models could become your IP. According to @lqiao, 90% of the world's data is still private and locked inside ap

JEPA world models + Hierarchical Planning is a massive step for long-horizon robotics. A classic failure mode I’ve faced with planning with world models: flat planning often 'cheats.' For example, in

Benchmarks are becoming the core competency of any company. Our company mission is to build the world’s best models for document processing. This also means that we have to build the world’s best benc

A gap that has been fixed in our methodology that highlighted a decline was not including local models (GGUF, AWQ, ONNX, etc). Earlier, these were far harder to track and closer to a rounding error. N

We're releasing pplx-embed-v2-late, two late-interaction embedding models that retrieve text, images, and pages with a shared embedding space for cross-model querying. Both models achieve frontier per

In case you missed it, here's what shipped last week: - Build with new models: GPT-6.1 Sol and Claude Sonnet 5.5 - Ask agent to use Jev via AI integration - Updated UI in Settings - Enterprise: Set ad

🪽DFlash draft models are available for Ornith-1.5 9B, 35B-A3B, and 397B. DFlash and Ornith 1.5 make a dynamic duo: DFlash performs as the draft model, accelerating inference by drafting multiple token

System one models are extremely useful for a variety of document tasks that require fast decisions: * orientation detection * language detection * classification * splitting We benchmarked Jev with ot

Simon Willison defines “Fable class models” as a lineage of generative AI systems that began with Claude Fable 5, progressed to Claude Opus 5.5, and later included GPT‑Astra 6 (with a possible additio

An interesting idea is to continually train specialized models on skills. This paper explores that idea. They propose SkillGym which turns human-written skills into 2,756 training environments with co

Recent Bloomberg reports indicate that a growing number of startups are developing custom AI models by leveraging open-weight alternatives to high-cost proprietary platforms like OpenAI’s and Anthropi

Meet Wally. Our inference stack for open frontier models, built to be the fastest place to run them. Performance snapshot: GLM-5.3 Flash: 380 tok/s. GLM-5.3 Max: 790 tok/s Qwen3.8-27B: 485 tok/s DeepS

What I like about Jev: for years, large generalist models sucked almost all the oxygen out of AI. But there are huge opportunities in much more specialized and customized models, built for specific ta

After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI? I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of fr

You are what you eat and AI models especially so https://x.com/EMostaque/status/2099580512675262512?s=20 The Labs need to focus on 10x or 100x their effort on ALIGNMENT. Stop training the models on th

Astra is one of the best models we've tested at parsing complex documents with tables, at 93.2% on ParseBench. So it seems well suited for ad-hoc analysis over table-heavy documents e.g. in finance. T

ChatGPT Desktop (the Codex app) now supports integration with Ollama models, allowing users to run local LLMs within the application. To utilize this feature, simply download or upgrade to Ollama vers

Looking forward to this ecosystem with open models continuing to flourish and grow with NVIDIA and Hugging Face, and the continued partnership. Congrats @JensenHuang and @ClementDelangue! Exciting day

This paper does a good job of showing the promise and gaps of autonomous AI scientists. Big question is how much more advanced models close those gaps. Today, we're excited to share early progress on

// There Is No Neutral Harness // Great work discussing some of the issues in harness evaluation. Twelve open-weight models answer the same 3,679 items from ARC, HellaSwag, MMLU, and TruthfulQA under

Top open models by use case 👇 chat - Kimi K3, Qwen3.8 2.4T A95B reasoning - Kimi K3, DeepSeek V4 Pro coding agents - Kimi K3, DeepSeek V4 Flash, GLM 5.2 small + fast - Qwen3.8 27B, gpt-oss 20B mid-siz

It is not inherently bad to publish research on the impact of AI that only refers to older models, but it requires a very careful discussion and has to be very clear to non-technical readers. If you s

GLM-5.3 (Max) by @Zai_org has shifted the Pareto frontier for Code Arena: WebDev! GLM-5.3 (Max) would sit at #2 among open models (upon open weights release, and #8 overall) with 1597 pts in the Code

Sovereign AI models lagging behind cutting‑edge systems risk significant disadvantages when applied to frontier tasks such as cyber defence. Ethan Mollick highlights that this penalty is most pronounc

Connect your coding agents to AI Gateway with a single command. • Auto-configure 8 popular coding harnesses • 300+ models from 30+ providers, no markup • Open-weight models with ZDR & US inference ▲ ~

imo GDPVal is probably the most important benchmark, it measures the performance of models on real world tasks Big leap in performance here to top it at a great price, congrats to @SpaceXAI team & loo

When should you start post-training your own models? @FireworksAI_HQ CEO @lqiao’s answer: after product-market fit. Not because it's hard... but because only after PMF is the data coming off your prod

I’ve been thinking about how agents can learn inside world models for years. We decided to scale up our RSI Lab to bridge recursive self-improvement with physical AI and robotics. We are looking for f

Open models give teams more room to run agent loops, with API prices at a fraction of GPT-5.6 Sol and Claude Fable 5. That matters as planning, tool calls, retries, and long contexts compound token us

This definitely seems like something worth noting, and illustrates the gap between Fable/Astra class models and the previous frontier that was 'merely' good at hacking under human instructions. Initia

The apps you use swap AI models constantly. Doing that without downtime or a bad rollout reaching users is still often manual. We rebuilt that workflow based on what we’ve learned serving more than 40

Very interesting paper on LLM reasoning. They find that frontier models can exhibit invisible reasoning by leveraging semantically irrelevant filler tokens. In other words, invisible reasoning can ser

Kimi K3 is now available for inference & training in @FireworksAI_HQ. Crazy how easy they make it to tune frontier open models like K3 using LoRA adapters. Best time to figure out how to own your inte

Open models are already being used in the enterprise. Over 85% of the Fortune 500 companies already use Ollama to fulfill specific tasks. @jmorgan Why open models and Ollama? Ownership. Open models ar

Generic AI models are not built for medicine. Doximity Ask, trained and served on Fireworks, outperformed GPT-5.6 Sol, Claude Fable 5, and OpenEvidence in a Stanford-Harvard clinical AI safety study.

Most models run on whatever inference engine they ship with. That's rarely the fastest option. We built our own, AGI-RUN. On a phone it beat every model's own native engine, and ran up to 9.9x faster

And the idea that frontier open weights models will continue to be released was fragile even before the current de facto licensing regime. https://x.com/emollick/status/2067669551685218638?s=20 Is the

VCs are now sharing screenshots in group chats of Claude discouraging investment in open-source AI infra startups and models. Obviously there is an absolute EXPLOSION of pitches in inference companies

The next scaling law is multi-agent swarms. Mixture of models. Introducing Sakana Fugu: A full multi-agent orchestration system accessible via a single model API. Our ‘Fugu Ultra’ model matches the pe