Strag retweeted
Wir verlosen eine NVIDIA GeForce RTX 5090 im individuellen Hogwarts Legacy-Design, um den offiziellen Release mit DLSS 5 zu feiern 🪄
So machst du mit:
🟢 Teile diesen Beitrag
🟢 Kommentiere mit #RTXON
Strag retweeted
We're giving away a Hogwarts Legacy custom wrapped NVIDIA GeForce RTX 5090 to celebrate its official release with DLSS 5 🪄
To enter:
🟢 Share this post
🟢 Comment #RTXON
Terms and Conditions: nvidia.com/en-us/geforce/con…
three people behind Claude Code just left Anthropic.
Claude Code is what made Anthropic the default for builders.
now i want to see where these three land, because nobody knows better what to build against it
Fable 5.5 just made a Ghibli short out of code.
JUST RAW CODE.
if models can one-shot animation like this in code, i don't know what video tools look like a year from now
this is the reason why i switched from Hermes to Grok Bot.
i still use Hermes, but most of my tasks have already moved over to Grok Bot.
what i love about Hermes: it's open source and you can use any model you want, including local ones. you can modify almost everything about how your agent works, and the community ships updates nonstop. if you love tinkering, it's hard to beat.
what wore me down: i was the one keeping it running, fixing the setup when something broke, and watching token costs. that's fine when building the agent is the fun part. it's less fun when you just want the work done.
Grok Bot took all of that off my plate.
i work on a desktop, a laptop and my phone, and it's the same bots, chats and tools on all three. i start something on one device and pick it up on another without setting anything up again.
my bots share their own cloud computer with a browser. they install tools, run code, keep files and stay logged into sites, and they keep working when my laptop is closed. computer use is fast too. it clicks through sites and comes back with the result.
i'm running a whole team of bots now, each with one job. they hand work to each other, remember what i've told them across chats, and save repeatable steps as skills so they do things the same way every time.
it's not perfect. you can't pick the model or self-host it, and the usage limits are real. if full control matters most to you, Hermes is still the better pick.
but for getting real work done every day across all my devices, Grok Bot is where most of my stuff lives now.
the main reason i love Grok Bot is how easy it is to switch devices. desktop, laptop, phone, it all feels the same. same tools, same features, same everything.
i love Claude Code too, but even with cloud sessions it doesn't feel the same when i jump to my phone. and i don't want to keep multiple devices open just to use remote control.
is there a better way to use Claude Code that i'm missing?
NVIDIA just dropped an official NVFP4 quant of DeepSeek v4.1 Flash
81 GB smaller and the quality is almost identical 👀
local AI keeps getting better and better xitter.cf/MiaAI_lab/status/21163…
some sources say Gemini 4 Xenon is next after Carbon
supposedly it beats the unreleased Fable 5.5, maxes out most benchmarks, and is way cheaper to run
google is cooking nonstop 😭
if even half of this is true, the rest of the competition is in trouble.
do you think google is capable of this? 👀
Grok 4.8 is coming, and Elon is promising it won't disappoint
the bar is Opus 5.5 on real work. the accessibility of grok bot + an Opus 5.5 level model would destroy the competition
i hope they release it this november.
gemini now lets you share your avatar and drop it into images and videos with friends
character and voice consistency built in. this is the feature every ai avatar creator has been waiting for. will definitely dive deep into this.
these Fable SVGs are clean enough to ship as real assets
a new Opus is coming too. is it finally time for Opus 5.5 to get mogged by its own sibling?
Anthropic is cooking
claude's mobile app just got a stack of real upgrades. multiple accounts, carplay voice, laptop mode on android, 33 languages
multiple account feature + seamless switching is definitely my favorite. it's probably close to getting shipped on desktop too 👀
GPT Image 2.6 is already looking like a big jump. the left version is the first time i've seen an image model treat 10 styles as 10 different worlds, not 10 filters on the same face
left is the "sunburst" checkpoint, right is "shooting-star". both were tested on arena before 2.6 showed up in openai's docs
credits to @chetaslua for the side by side images.
the #1 model in the world right now and we can't use it for another month? 😭
are they waiting to get mogged by Fable 5.5 before they release it?
I did say we'd be releasing our models periodically...introducing Gemini 4 Carbon, our next frontier model.
Following Argon's release just two weeks ago, Carbon shows the pace at which we’re pushing the frontier, it ranks #1 on the recently updated v5.0 of @ArtificialAnlys Intelligence Index at 59. It also leads across key cybersecurity evaluations scoring 88.2% on SEC-bench Pro for finding vulnerabilities, 54.8% on ExploitGym for exploitation, and 71.0% on CWE-bench for patching.
We’re seeing significant improvements in long-horizon tasks even compared to Argon, and we’ve strengthened frontier safeguards.
As we did with Argon, we’re going to roll it out gradually + responsibly. Some trusted cyber defenders will get access starting this week. We'll be expanding it to more partners through our Fairwind Program, ahead of wider availability soon, so hang tight!
doubled limits is a W. but limits only matter if the model is worth using
your move @OpenAI, show us something better than Opus 5.5
Day 25/
We are doubling all usage limits for all Pro 100/200/500 tiers and Business Premium, effective today.
We heard your kind feedback and won’t be reducing usage limits for Pro 200, so your Work and Codex usage will stay the same. We’re also going to continue offering extended limits for dots for now, and usage of your dot will still continue to be separate from your default usage.
And we'll of course also keep shipping better models and other efficiency improvements that increase your effective usage over time.
le chonk is the first open model outside china that feels like a real option
Mistral Large 4 is now open weights, narrowly making it the #1 open weights model developed outside of China
@MistralAI has updated and released the weights of their flagship model, Mistral Large 4, with 1.05T total and 52B active parameters. With high reasoning, Mistral Large 4 scores 26 on the Artificial Analysis Intelligence Index v5, up 1 point from Mistral Large 4 Preview (25), with the major differences being increased token efficiency and reduced hallucinations. It sits 1 point ahead of the leading US-developed open weights model, Muse Spark Open (xhigh, 25).
On the Artificial Analysis Cyber Index, Mistral Large 4 (high) scores 51, third among current open weights models behind DeepSeek V4.1 Flash 1017 (max, 57) and MiMo-V2.6-Pro (56).
Key takeaways:
➤ Narrowly leads non-Chinese open weights models: Mistral Large 4 (high) leads Muse Spark Open (xhigh), the next-best open weights model developed outside China, with the largest gap on Terminal-Bench 4.0 (25% vs. 8%). It trails leading open weights models DeepSeek V4.1 Flash 1017 (max) and GLM-5.3 (max) by 6 points on the Intelligence Index, largely due to weaker coding and agentic knowledge work performance: 15% on FrontierCode 2.0 (vs. 29% and 22%, respectively) and 1349 Elo on AA-Briefcase (vs. 1570 and 1504)
➤ Increased token efficiency: Mistral Large 4 (high) uses 82K output tokens per Intelligence Index task, 39% fewer than Preview (135K), placing it on the open weights Pareto frontier. The seven open weights models that score higher use between 113K and 225K output tokens per task
➤ #2 open weights model on the Coding Agent Index: On the Artificial Analysis Coding Agent Index, Mistral Large 4 (high) scores 48 in Mistral Vibe while using 8.5M tokens per task, about half as many as the leading open weights model GLM-5.3 (max) in Opencode (49, 15.9M). It also scores above another leading open weights model, DeepSeek V4.1 Flash 1017 (max), in DeepSeek Harness (46, 16.7M)
➤ Reduced hallucinations: Mistral Large 4 (high) has a 28% hallucination rate on the updated AA-Omniscience v1.1, down from 41% for Preview and matching Muse Spark Open (xhigh). However, its accuracy remains relatively low at 26%, similar to Ling 3.1 Flash 1019 (27%), and behind DeepSeek V4.1 Flash 1017 (max, 40%) and Muse Spark Open (xhigh, 42%)
➤ Lower price and cost per task than the Preview: Mistral Large 4 is priced at $0.70/$2.10 per 1M input/output tokens, a ~50% reduction from the Preview's pricing of $1.36/$4.18. At the new pricing, it costs $2.06 per Intelligence Index task, down 63% from $5.54 for the Preview. This is comparable to Muse Spark Open (xhigh, $2.02) but more costly than leading open weights models DeepSeek V4.1 Flash 1017 (max, $0.60) and MiMo-V2.6-Pro ($0.26)
Key model details:
➤ Size: 1.05T total parameters / 52B active parameters
➤ Context window: 1M tokens
➤ Multimodality: Text and image input, text output
➤ Pricing: $0.70/$2.10 per 1M input/output tokens ($0.07 per 1M cached input tokens)
➤ License: MEOW License (Mistral Ecosystem Open-Weight License), where operating a Model-as-a-Service business requires a separate commercial license with Mistral
Congratulations to the team at @MistralAI on the release!
learned this the slow way. fixing the output without fixing the instructions just means fixing the same thing again next week.
I feel stupid saying this, but you gotta make sure your agents don't repeat the same mistake.
Every time you find an issue that goes through your agentic review, ask the following:
• What specific check should have caught this bug?
• At what stage should that check have run?
• Why didn't the existing tests, rules, or instructions catch it?
• What can we change so this problem gets caught automatically?
Then:
• If it was a missing test, write one.
• If it was a missing review rule, add it.
• If the instructions were unclear, improve them.
• If the agent was missing context, provide it.
I know it's tempting to fix the issue and move on, but this is not smart.
Take a minute to fix the process that let the problem happen in the first place.