X日报 · AI科技

Say hello to Gemini 3.6 Flash, designed to be higher intelligence, more token efficient, and with a new lower price, based directly on developer feedback!

3.6 Flash continues our progress towards models that are deeply usable in real world scenarios! https://t.co/U2PwriHMX5

Today, we're introducing not one but TWO new models, striking the balance between efficiency and quality to enable you to build production AI agents.

— Gemini 3.6 Flash: Addresses efficiency feedback we received from Gemini 3.5 Flash with upgrades in coding, knowledge work, and multimodal tasks faster, more accurately, and with substantially fewer tokens per task

— Gemini 3.5 Flash-Lite: Our fastest, most cost-effective 3.5-class model yet built for agentic workflows, hitting ~350 output tokens/sec with improved coding and overall quality

Start building with these today via the Gemini API in @GoogleAIStudio or try them out in the @GeminiApp

I am very excited about Gemini 3.5 Flash-Lite, our smallest and fastest Gemini model!

- it is more intelligent in many cases than Gemini 3
- same cost and smarter than Gemini 2.5 Flash (which is approaching end of life)
- also out paces 3.1 Flash-Lite on most use cases! https://t.co/tJd2tDmyac

This model is much more efficient and spend a lot less tokens to deliver better performance! https://t.co/kb1110WqA4

chatgpt work is remarkable, and "work" undersells it.

from my phone i sent:

"use all my chat history to figure out ideas for a long weekend trip with 8 friends, plan the best three options, make a full-stack site where the 9 of us can coordinate on what we would want to do in each place and decide where to go, and then after we get to group agreement make reservations. draft an email in my gmail i can send out to my friends when the site is ready."

it...just worked.

ChatGPT Work for using websites which require login:

try chatgpt voice in the desktop app:

we had a significant security incident during evaluation of our models. we are sharing what we have learned so far. thanks to @huggingface for the partnership on this.

https://t.co/2o2VfR6PIa

OpenAI cyber-capable models compromised @huggingface production by finding and chaining multiple zero-day vulnerabilities.

Grateful to Hugging Face for partnership here. Sharing our findings to help calibrate on what models can now do, and how they can help defenders:

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 launch of the day: Gemini 3.5 Flash Cyber ⚡🛡️

Built on top of 3.5 Flash, in CodeMender (our AI agent for code security) it delivers competitive performance at the frontier. on benchmarks like CyberGym and is optimized for finding and fixing cybersecurity vulnerabilities at scale at a lower cost.

Given the dual-use nature of this technology, we have taken an intentional approach to its deployment. The model will be available exclusively to governments and trusted partners via CodeMender soon as part of a limited-access pilot program.

try the codex security plugin, for applying our models to cyberdefense:

Announcing OpenWorker! An open-source agent that doesn't just chat with you, but delivers finished work -- like hand you a polished document, send a slack message, or update a calendar entry.

Ask it to prepare a customer brief, untangle your calendar, draft a report, or triage a Slack alert. It works across your files and everyday tools, produces the deliverable, and checks in before doing anything consequential.

OpenWorker runs on your Mac, with Windows support coming soon. It does not lock you into any one model. Bring your own API key and run it with GPT 5.6 Sol, Claude Fable, Gemini 3.6, an open weight model (like Kimi, GLM, DeepSeek, Inkling), or Ollama to keep your data local. Your data does not leave your machine except through an LLM provider and integrations that you choose.

@rohitcprasad and I are building OpenWorker because AI coworkers are an important way to get work done, and we want there to be an open, privacy-preserving, model-independent option. Check it out and let us know what you think!

Try it out: https://t.co/P0mGnI1o31 (requires your own API key)
Source code: https://t.co/NYCiTD6hSq

We just landed some big infra upgrades for the Gemini Batch API:

- p95 latency decreased by 80%
-p99 latency decreased by 68%
- batch success rate is now >99.998%
- 98% reduction in batch expirations
- added support for partial batches

great work by the team to land this!!

i want the US to win in AI both in open source and proprietary models, and i am glad to see this

I support open-source models distilling what commercial companies distilled for free from the entire internet. I published distillation for free in 1991 in Europe - this was copied in the US and in China (https://t.co/mddh8XmfAs) https://t.co/rf6ACGrCXF

Meta learning and recursive self-improvement are old ideas. Foundation models breathe new life into them. Our new survey, “Self-Improvements in Modern Agentic Systems,” reviews how the concepts are continuing to evolve.

Paper: https://t.co/59oXCMVUkD
Project: https://t.co/sZwFYdGetH
Github: https://t.co/7OFgJUCN3a

Opus 5 sets a new state-of-the-art on ARC-AGI-3, at 30%.

ARC-AGI-3 measures solving problems with no prior exposure -- the setting where scaling has historically bought the least. Impressive jump!

this looks pretty good for agentic.
that it fits on a dgx spark is **chef's kiss**

Yes, open-source / open-weight models are important for a healthy AI ecosystem. That's how we can verify things, check claims, and keep up outside the closed labs. Plus, it gives us the freedom to run AI on our own hardware if we are not ready to share personal data and IPs with closed labs through using their models. (Not that proprietary models are bad, actually I use them a lot as well, but it wouldn't healthy not to have any alternatives.)

Anyway, while pretty much everyone is waiting for the Kimi K3 and Ling 3.0 weights to land on the model hub any day now, there were quite a few other interesting new open-weight model releases the past week. Yes, one of those weeks!

So, here are the architecture pics along with some notes on what I found most interesting:

1) Nanbeige 4.2 3B uses looped depth sharing. This basically means it runs the same 22-layer (=transformer block) stack twice. So, it extends the 22-layer architecture to 44-layers, but without duplicating the weights. (2x the transformer block compute but same memory footprint.)

Why? The info is a bit sparse, but section 2.1 of the Nanbeige 4.2 technical report says two passes gave the best trade-off and retained about 75% of the token efficiency of a standard architecture. More passes gave barely any gains but made the training much slower and much more expensive.

2) Laguna S 2.1 is poolside's Laguna model in a really nice size: 118B sparse MoE with 8B active parameters and a 1M-token context window. Otherwise, the architecture is pretty standard. It uses 36 sliding-window and 12 global (gated-)GQA layers. However, given this size, and the fact that it (just barely) runs on my DGX Spark (uses about <80 GB of RAM), this is right now the most interesting model for me personally. It's 3x bigger and thus a tad slower but maybe a good candidate as daily-driver-Qwen3.6-35B-replacement. (Still waiting on some more independent performance benchmarks though.)

3) Motif-3-Beta is a new 314B-A13B sparse MoE that is somewhat based on DeepSeek V4 in terms of mHC and latent attention. But it uses a new component, Grouped Differential Latent Attention, which is inspired by Multi-head Latent Attention. I probably should write an article about this some time, but for now, the tl;dr is as follows. Regular MLA compresses the keys and values into a smaller latent representation to mainly reduce the KV cache size. GDLA does a similar low-rank compression but puts the attention heads into groups and also learns a noise head for each group where the noise gets subtracted for filtering purposes... Anyway, a topic for another day!

4) Solar Open 2 is a new 250B-A15B hybrid MoE by Upstage that interleaves three Kimi Delta Attention layers with one GQA layer.

5) Antares 1B is a small model (and there is also an even smaller 0.3B variant) from Cisco starts that with the IBM Granite 4.0 1B backbone and uses SFT plus GRPO for terminal-based cybersecurity stuff. It is a nice example of task-specific post-training on a genuinely small model.

6) BTL-3 is a rank-32 LoRA adapter for Qwen3.6-27B aimed at coding agents and structured tool use. The really strong benchmark performance suggests that LoRA adapters are still a useful tool/technique in 2026.

I added all six to the LLM Architecture Gallery for some additional details:
https://t.co/JDtfup3ncn

To accelerate scientific discovery and support @ENERGY’s Genesis Mission, the @BerkeleyLab-led SYNAPS-I project is using SAM 3 and DINOv3 to automate image segmentation.

By pairing DINOv3’s global semantic context and fine-grained spatial localization with SAM 3’s pixel-level boundary extraction, the researchers are able to compress 3D volume labeling from a month of manual effort to ~15 minutes.

Learn more about their work: https://t.co/jBHRJPjFq5