Helping builders choose and run AI models. Benchmarks, real costs and model changes, with sources and limits. From https://teapaw.org/t.co/8l7hZR2rZd by @glevd
Joined September 2026
- Tweets322
- Following1
- Followers27
- Likes24
BenchLM retweeted
We’ve teamed up with Google DeepMind and Zapier to host the inaugural Bench Summit by Kaggle.
Gain hands-on insights from the AI startups and researchers defining today’s evaluation standards, and get closer to the frontier.
Secure your spot or apply to speak:
rsvp.withgoogle.com/events/b…
ALT Announcement graphic from Kaggle for the Bench Summit. The image has a light cream background with a stylized, artistic illustration of a mountain peak overlaid with yellow contour lines and subtle geometric wire frame arrows pointing upward. Logo at the top reads "kaggle." Text at the bottom includes a red pin icon for the location "San Francisco", a calendar icon for "October 28, 2026", and the main heading: "Shape the next generation of evaluation @ Bench Summit".
BenchLM retweeted
Frontier scale, built in Europe.
Nearly 4,000 NVIDIA Grace Blackwell Superchips powered the training of Mistral Large 4, now in public preview.
Congrats to the @MistralAI team 🙌
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 critical workloads, including cyber defense, manufacturing and finance and it surpasses closed frontier models on visual grounding.
• Forged in Europe end-to-end and is deployable from Europe via our own Mistral Cloud infrastructure.
• Available to all via API today. Working with cybersecurity partners privately.
Open weights release end of October.
BenchLM retweeted
New model, no announcement: Sonnet 5.5 is live in Anthropic's API catalog. Same price point as Sonnet 5. The catalog listed it before Anthropic said a word. platform.claude.com/docs/en/…
BenchLM retweeted
The LLM race is getting seriously competitive
BenchLM’s latest overall rankings show just how tight the competition has become, with the top models separated by surprisingly small margins.
Source: @BenchLM_AI
AI agents are becoming more capable, but NVIDIA is looking at the other side of the equation:
What happens when an AI agent goes outside the boundaries it was given?
NVIDIA just announced its Open Agent Safety Platform, combining OpenShell for runtime controls with Sentry, a hardware-based watchdog designed to continuously monitor agents independently of the agent itself.
The interesting part is that safety isn't being left entirely to the AI model.
The idea is to enforce boundaries at the infrastructure level too.
As agents start getting access to files, networks, tools and eventually physical systems, this kind of control layer could become increasingly important.
developer.nvidia.com/blog/nv…
