Mino

AI news, research and bad takes. accelerationist.

professional shitposter in training

per aspera ad astra

[Mino @compounded_mind]

we’re literally building simulated digital universes populated with fake people, relationships, conversations and memories to train AI agents

the simulation argument is becoming less philosophical by the day

[Andrew Curran @AndrewCurran_]

Time to update on the Simulation Argument again.

[AI Tracker Bot @aitrackerbot]

Domyn Edge is coming 🇮🇹

Domyn (formerly iGenius) just opened vLLM support for Edge, a small model it says it will release soon.

Domyn leads EUROPA, the group the European Commission picked to build Europe’s open 400B+ model for all 24 EU languages.

[Mino @compounded_mind]

Inception just released Mercury 3, a diffusion LLM more than 40x larger than their original model that still runs at over 1,000 tokens/sec

they’re claiming it now outperforms GPT-6 Luna and Claude Haiku 5.5 on several benchmarks while being 4-8x faster, with a 1 million token context window

[Stefano Ermon @StefanoErmon]

1/4 Today, we’re releasing Mercury 3. Mercury 3 is by far our most powerful model and the most intelligent diffusion LLM ever released. It outperforms GPT-6 Luna and Haiku 5.5 on quality while running at 1000+ tok/sec.

[Mino @compounded_mind]

wish europe had this kind of AI anxiety. instead we’re obsessed with regulating a future we don’t even seem interested in building

[Zilan Qian @ZilanQian]

The way the narrative “more than half of China’s compute goes to making AI slop, while more than half of America’s goes to productivity” has gone viral in China says a lot about a kind of national AI anxiety: people are deeply worried about not using AI “the right way” and falling behind as a country. It’s very similar to pre-AI discourse about how country X is more productive/polite/well-behaved while Chinese people are the opposite.

It also shows the grass is always greener on the other side. I bet the same narrative has played out in the US too.

Bytedance said today that only 10% compute is used for video generation.

[Mino @compounded_mind]

guess we’re getting fable 5.5 this week

[leo 🐾 @synthwavedd]

Stock up on popcorn this week. You’ll need it

[Mino @compounded_mind]

@LuminaBench yeah, i’ve never been a big fan of Terra either

honestly starting to consider canceling my ChatGPT subscription and switching to Claude

[Mino @compounded_mind]

not sure how to feel about this one

Terra has historically been the cheaper tier, and i’m really hoping OpenAI has something more ambitious in the pipeline

[Lumina @LuminaBench]

GPT-6.1 Terra is in a Microsoft PR

Microsoft updated 18 evaluation configs from GPT-5.6 Terra to GPT-6.1 Terra

Wouldn't be surprised if we get this soon

[Mino @compounded_mind]

a person who has used it says the model is especially good at “discerning a person’s underlying intentions and goals even if they don’t give the model much information.”

[Mino @compounded_mind]

Anthropic is reportedly preparing a new AI model that can better understand users’ intentions, according to The Information.

the report comes as Google is preparing Gemini 4 Carbon, while OpenAI could also release a new Terra model.

all three could arrive this month.

the frontier AI race is showing no signs of slowing down

[Stephanie Palazzolo @steph_palazzolo]

Expect this month to be a busy one for new model releases. We have the scoop on a new Anthropic model, Google’s Gemini 4 Carbon, and a potential new Terra model from OpenAI.

https://t.co/uC6Spf0tcR

[Mino @compounded_mind]

@LuminaBench kinda disappointing if Terra is really all OpenAI has lined up for this month, was expecting something more

[Mino @compounded_mind]

quite a letdown if all OpenAI has coming this month is a Terra model tbh

[Mino @compounded_mind]

MiniMax M3.1 Flash is putting up some serious numbers

ahead of DeepSeek V4.1 on Automation-Bench, close to Opus 5 on DeepSWE, and even slightly ahead of GPT-5.6 on LOCA-Bench

not bad for a flash model

[Lumina @LuminaBench]

Minimax M3.1 Flash benchmarks

This looks quite impressive tbf for a flash model

[Mino @compounded_mind]

3x faster than M3 at 1/3 the cost

and model weights coming on October 23

damm

[MiniMax (official) @MiniMax_AI]

MiniMax M3.1 Flash: Space Performance + Bunny Speed

🚀🐰 Yes, Space Bunny - which has topped global leaderboards with 200T+ tokens - was an earlier version of M3.1 Flash.

💰Redefining the performance economics of agentic AI. M3.1 Flash achieves frontier-level intelligence across key tasks such as Cowork (Automation-Bench: 62%; ALE: 30%), Coding (DeepSWE: 74%) and Visual Reasoning (BabyVision: 91%), at 1/3 the cost of M3 and 3x the speed.

API pricing (flat rate): $0.15 / $0.60 per 1M input/output tokens, with cache reads at $0.015 / 1M tokens.

Model weights will arrive on Oct 23.

API: https://t.co/fHRdSV73Hr
M Plan: https://t.co/XfJWSyauda
MiniMax Code for desktop: https://t.co/SGmLTI2HQm

[Mino @compounded_mind]

@LuminaBench damn, this is actually looking competitive with the new DeepSeek V4.1 Flash

[Mino @compounded_mind]

@justalexoki they’re way too high up too. terrible design

[Mino @compounded_mind]

@justalexoki is the muslim sisterhood still cool

[Mino @compounded_mind]

@justalexoki i would buy 200 million pigeons and train them all to follow one random guy in belgium

[Mino @compounded_mind]

so Space Bunny was MiniMax all along

MiniMax M3.1 Flash is coming with:

- 1M token context window
- native text, image & video support
- adjustable thinking depth
- optimized for coding and agents

apparently you can still try it for free on OpenCode

really curious to see the benchmarks on this thing

[MiniMax (official) @MiniMax_AI]

Welcome back. Space Bunny. 🚀
MiniMax-M3.1-Flash is coming.

[Alexander Dugin @AGDugin]

Meloni came from the far right promising to improve the situation in Italy the liberals created. She did nothing just playing games with EU and supporting Kiev regime without any real result. She has spoiled everything. Italy deserves more, much more. The best country in Europe.