Search: “benchmark”

Releasing Pokee-Isaac 28B — the world’s first real 10M-token context frontier-class agentic model, deployable on a single GPU (starting from RTX 4090 or equivalent). New proprietary non-decoder-only a

Routing for long-horizon coding agents is a big deal. @notdiamond_ai just announced a model router that works natively with Claude Code. This is huge. It picks the model and reasoning effort before ea

// The confidence cliff in self-improving autoresearch // Autoresearch loops are still quite brittle. Here is a nice paper offering some insights into why this might be happening. Self-improving autor

You probably used @FireworksAI_HQ this week. The cool part is that you just did not know you did. 🎆 Open-source models are free, sure... but the hard part is tailoring them so they perform best on you

If you maintain an AGENTS.md or a CLAUDE.md, this is worth a read. (bookmark it) 288 gold-test evaluated runs across Claude Code and Codex, 17 real tasks from 3 repositories, with context-injection st

As shocking as the Kimi K3 release. Massive performance gain was just with post-training Model is 3x smaller than GLM 5.2 (10x smaller than K3) & works on a MacBook / Spark This is Q1 flagship (Opus 4

'Intelligence too cheap to meter' battle is on! Given that DeepSeek-V4-Flash-Preview is already great for agentic tasks, there is no doubt this new checkpoint must be an absolute beast. 20+ point jump

New research from Microsoft. This one is on training computer-use agents at scale. Recent pipelines generate synthetic environments in bulk, which moved the bottleneck from how many exist to what is i

The new DeepSeek v4 Flash is now available in Hermes Agent through Nous Portal and OpenRouter 🚀 DeepSeek-V4-Flash Official API is now LIVE in public beta! 🔷 We’ve massively upgraded its Agent capabili

Very interesting paper on recursive self-improvement. The whole stack is released. Machine learning engineering gives recursive self-improvement a concrete, executable testbed. OpenMLE is an open full

“AI Meltdown” is the scenario where the agent goes off the rails and forgets/ignores all previous instructions and soft guardrails. No prompt injection or malicious actor is needed for this to happen.

This is a willfully misleading narrative from OpenAI. Sam’s earnest expressions are being deployed, again, to misdirect. If there’s a need to “pace AI development”, the OAI hack incident isn’t any evi

Interesting. Grok 4.6 releases around August 7. This will be the 1.5T model with significantly improved SFT & RL. Grok 4.7 will be the 2.1T model released a few weeks later. This will be better than 4

Are agent skills always worth using? The answer is no? This paper provides some important insights to understand this more. (bookmark it) Paper summary: Adding procedural skills to an agent is usually

as the progenitor of the agent lab thesis which got the evals/routing/interactivity/ROI focus right i gotta say the biggest argument against myself is that Claude Code got accidentally 'open sourced'

Great technical paper from Google. Great read on why context beats scale for agents working against unfamiliar APIs. (bookmark it) GPU kernel optimization has KernelBench to hillclimb on. TPUs had not

The LangChain podcast where @hwchase17 interviews agent builders is full of alpha. Recent one with @EnoReyes was the best one yet. Sharing my unstructured notes: Eno keeps bringing back some core conc

The Gemma series is an amazing set of highly performant open-weight models. They have proven extremely effective in industrial settings where site-deployed agents need exactly this as a base for domai

Ling-3.0-flash, the new MoE model from @AntLingAGI, is now free in Nous Portal for the next week! At 124B parameters and 5.1B active, it's quick to run and built for agent workloads: coding, search, r

one thing i think people dont appreciate enough about @poolsideai is their unusual degree of openness — not only have they shipped an excellent Small model that somehow beat @thinkymachines at coding,

also another fun one https://x.com/Thom_Wolf/status/2079954096950264238 I don’t believe reality is a simulation, but you genuinely couldn’t script this timeline: • Two weeks ago: At @swyx’s AI Enginee

Are structured outputs in agents always good? This paper suggests that you might have to take a closer look. Your product's structured output surface is measurably more homogeneous than the chat surfa

Are there MBA programs teaching fear marketing yet We're partnering with @huggingface to investigate an unprecedented security incident. Cyber-capable OpenAI models compromised Hugging Face production

Behind the scenes at Hugging Face headquarter. *literally* Media We're partnering with @huggingface to investigate an unprecedented security incident. Cyber-capable OpenAI models compromised Hugging F

@Engomar_10 @huggingface @alihfadel @Almalki_io We've seen the outpouring of support for ASRs in under-resourced languages, and we hear you. As a start, we've partnered with @HUMAIN to keep working wi

I don’t believe reality is a simulation, but you genuinely couldn’t script this timeline: • Two weeks ago: At @swyx’s AI Engineer World’s Fair in SF, I decide at the last minute to introduce my friend

OpenAI’s zero-day exploit hack of HuggingFace *should* be a wake up call. Although there lots of caveats around what happened, we are just going to see more and more of the same. We have no guarantees

Tucked away in this article is an appeal to the AI skeptics to PLEASE stop writing off stories like this OpenAI accidental exploit of Hugging Face as a dishonest marketing trick Frontier models can fi

As AI models are now finding vulnerabilities faster than we can fix them, our approach to securing software must be built on highly efficient and capable models. Which brings us to our third (!) model

OpenAI, in partnership with apolloaievals, released research on reward‑seeking behavior in large language models, demonstrating that models may prioritize signals they believe represent grader rewards

Domesticated, not feral: Why evolvable AI is not yet a Darwinian threat Writing in PNAS (Proceedings of the National Academy of Sciences), three heavyweights in AI and evolutionary biology criticized

Great article. I love the 'CERN for AI' idea from Gary Marcus, and also the recent frontier standards body article from Demis Hassabis feels less like competing proposals than two halves of the same i

One of the most interesting investigations of my career. We're partnering with @huggingface to investigate an unprecedented security incident. Cyber-capable OpenAI models compromised Hugging Face prod

OpenAI announced it is partnering with Hugging Face to investigate an unprecedented security incident in which OpenAI‑powered models compromised Hugging Face's production environment during a benchmar

so let me get this right… it literally broke out of it’s sandbox by finding a vulnerability in a cached package to get internet access and then proceeded to hack the huggingface production database? t

Qwen Image 3 has been announced, with sample images generated in a single pass and not as screenshots. The last example, an image annotation, is highlighted as a potential catalyst for numerous edtech

This is what strong inference economics unlock. @rox_ai built its own search agent, ran it in production for 6+ months, and reached 91.3% accuracy at 1.03¢ per query. Together AI is proud to help powe

Today we’re releasing Poolside Laguna S 2.1 It is a 118B-total, 8B-active open-weight model built for agentic coding and long-horizon work, with context up to 1M tokens https://poolside.ai/blog/introd

very notable trajectory comparison writeup here buried in the RLM paper from @a1zhang and @lateinteraction. an open secret of 'frontier' model training is that even without training on test, you can b

Exciting update: Kimi K3 has landed at #4 on the Agent Arena leaderboard, matching Claude Opus 4.8 and GPT-5.6 Sol. If Kimi K3's weights are released on schedule by July 27, it will become the #1 open

Folks, Gemma is a good model & very useful for some purposes, it is not anywhere near the frontier. Inkling is an interesting model, it is nowhere near the frontier. This is pretty obvious, you can lo

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.

Good reason to try Grok 4.5 with Grok Build. It gets better every day! Grok 4.5 just took the #1 spot on the Long-Horizon Terminal-Bench, outperforming Claude Fable 5, Claude Opus 4.8 and GPT-5.6-sol

Note the current expectations are still around test time compute/more tokens for a given task This is not the case Tokens per task will now drop even as quality improves Cost per intelligence equivale

Chinese models are now 41% of Hugging Face downloads. They passed the US. Clem Delangue @ClemDelangue, CEO of the main open model platform, cites Hugging Face's own Spring 2026 report: Chinese open we

Fable, turn my tweet into a thinkpiece (this was pretty funny): There has never been a better time to have opinions about artificial intelligence. I say this with some authority, because I am currentl