iiinova — where your digital self keeps evolving

Sujee Maniyam @sujee.dev · 03/10/2026

Most AI models write. Jev decides. A 2-minute explainer for devs. Made with Claude Code + Opus 5.5. Open source and fully reproducible. Remix it: github.com/sujee/visual... #AI #LLM #ExplainerVideo #Jev

110

Sujee Maniyam @sujee.dev · 02/10/2026

Here’s the crinkled/scanned-looking document image (generated!) I fed into the model 👇 Not exactly a clean input - which makes the result even more fun.

000

Sujee Maniyam @sujee.dev · 02/10/2026

Quick demo showing the vision capabilities of Qwen 3.8-27B 👇 I gave it a crinkled, scanned-looking document containing PII. It extracted the text, analyzed the document, and identified the PII ✅ This little model is a beast 😄 And it was ticking along at 272 tok/sec on @nebiustf 🚀

100

Sujee Maniyam @sujee.dev · 25/09/2026

Try it for yourself 🐍 sujee.github.io/llm-snake-ar... Go burn some tokens :-) Code: github.com/sujee/llm-sn...

000

Sujee Maniyam @sujee.dev · 25/09/2026

Friday funsies 🐍 LLM Snake Arena: GPT-6-Astra vs Claude Opus 5.5 Does it tell us something about the models? Maybe. Fun to watch? Most definitely 😀

100

Sujee Maniyam @sujee.dev · 12/09/2026

Top open models now: 1. GLM-5.3: 45 2. Kimi K3: 44 (60 → 44 = ▼ 16) 3. GLM-5.3-Flash: 42 (57 → 42 = ▼ 15 ) Big score shifts, but the important thing is the relative ranking under the new benchmark suite.

000

Sujee Maniyam @sujee.dev · 12/09/2026

Artificial Analysis just updated its Intelligence Index to v4.3 with a tougher benchmark suite. Scores dropped quite a bit. Top proprietary models now: - Claude Fable 5.1: 53 = 66 → 53 = ▼ 13 - GPT-6 Astra: 53 = 61 → 53 = ▼ 8

210

Sujee Maniyam @sujee.dev · 07/09/2026

Went to buy a 4TB portable SSDs and almost fell off my chair looking at the prices 😳 Damn. See screenshots 👇 @demian_ai has been writing about the memory/storage crunch driven by the AI boom. It is finally sinking in for me :-) x.com/demian_ai/st... x.com/demian_ai/st... x.com/demian_ai/st...

000

Sujee Maniyam @sujee.dev · 07/09/2026

Fine-tuning, distillation, and quantization are easy to blur together. Here is an explainer video with LEGO® bricks 🧱 🟢 Fine-tune = specialize 🟠 Distill = teach a student 🟡 Quantize = use lower precision More detail: sujee.dev/post/fine-tu... drop your suggestions in comments

000

Sujee Maniyam @sujee.dev · 29/08/2026

2️⃣ Second pic: GLM-5.3-Flash running inside OpenCode. Amazing speed and performance! ⚡

000

Sujee Maniyam @sujee.dev · 29/08/2026

1️⃣ First graphic: GLM-5.3-Flash (formerly Ox Alpha!) scores 57 on the Artificial Analysis Intelligence Index, while Kimi K3 scores 60. But GLM-5.3-Flash is: much cheaper, much smaller Try it for yourself : tokenfactory.nebius.com/playground?m... Model visualizer: sujee.github.io/practical-ll...

100

Sujee Maniyam @sujee.dev · 28/08/2026

I tested 3 NVIDIA Nemotron models with the same sales-analysis agent. All scored 100%. Lightning done in 7s for $0.0019 - about 7x cheaper than Super and 16x cheaper than Ultra. TLDR; For focused agent workflows, smaller may be enough. www.youtube.com/watch?v=Ycme... github.com/nebius/token...

000

Sujee Maniyam @sujee.dev · 07/08/2026

Kimi K3 was one of the most anticipated open models. Join the Kimi + Nebius Token Factory teams on Aug 12: 🏗️ K3 architecture + capabilities ⚡ Inference engineering + optimizations 🎁 Token Factory credits to try K3 yourself 9 AM PT / 12 PM ET / 6 PM CET 👉 nebius.com/events/webin...

000

Sujee Maniyam @sujee.dev · 29/07/2026

Trying out @kimi-moonshot-x.bsky.social K3 (running on Nebius Token Factory ) for coding with @opencode.bsky.social . The integration was incredibly smooth - easy to configure and start building. Here’s a quick 30-second walkthrough to get you up and running. 👇 start: tokenfactory.nebius.com

110

Sujee Maniyam @sujee.dev · 23/07/2026

Nebius Token Factory @nebiustf Office Hour #3 Recap: We covered new models and features, inference optimizations, fine-tuning, the Builder and Fellow programs, plus live Q&A. recording + slides in comment 👇 See you at the August office hour: nebius.com/events/webin...

100

Sujee Maniyam @sujee.dev · 22/07/2026

Building with open models? Join us for Nebius Token Factory Office Hour - a live, developer-focused session featuring demos and Q&A 🗓️ July 22, 2026 ⏰ 9 AM PT / 12 PM ET You’ll also receive Token Factory credits, so you can start building right away. See you there! nebius.com/events/webin...

000

Sujee Maniyam @sujee.dev · 02/07/2026

Nebius booths are never boring. Meet NEBU - our Nebius pup 🐶 - making a new friend 🐉 Stop by the Nebius booth at AI Engineer World Fair. You can play with the robots... and chat with the humans, too. 😄

000

Sujee Maniyam @sujee.dev · 01/07/2026

Come by the Nebius booth at AI Engineer conf! We’ve got fun demos: LLMs playing Snake 🐍, LLMs playing soccer/fútbol ⚽️ Demo below 👇 We’ll also show you our open-model coding stack - GLM 5.2 and Kimi K2.6 are our current faves. Plus, we’ll set you up with Token Factory credits 🎁

000

Sujee Maniyam @sujee.dev · 24/06/2026

I’m running a workshop at ClawCamp Campout @ Dual Tech Summit! We’ll add real-time intelligence to OpenClaw using Nebius Token Factory + Tavily 🗓️ Wed, June 24, 2026 📍 War Memorial Veterans Building, SF 👉 clawcamp.us/event/260624... 🎁 Free Token Factory + Tavily credits . See you there!

110

Sujee Maniyam @sujee.dev · 19/02/2026

The AgentBeats–AgentX Competition Phase 1 winners have been announced - and the projects are fantastic. Huge congratulations to all the teams involved! 👏 Read the highlights here: berkeleyrdi.substack.com/i/188179396/... Now, on to Phase 2 : rdi.berkeley.edu/agentx-agent...

000

Sujee Maniyam @sujee.dev · 18/02/2026

(another car wash post) Q: I want to wash my car. The car wash is 50 meters away. Should I walk or drive? Kimi K2.5: (thought 52 secs) : Drive the 50 meters, but maybe take the long way around the block to let the engine warm up slightly ...

100

Sujee Maniyam @sujee.dev · 13/02/2026

Kimi 2.5 from @kimi-moonshot-x.bsky.social is now live on Nebius Token Factory. A major step up from Kimi 2, this os multimodal model is strong in agentic workflows and coding tasks - and it’s stacking up well against SOTA proprietary models. Check out the official benchmarks and the pelican 🪿🚲

100

Sujee Maniyam @sujee.dev · 15/01/2026

MiniMax M2.1 from MiniMax is now live on Nebius Token Factory. strong showing for coding, agentic workflows It’s also ranked among the **top open models** by Artificial Analysis And yes… it passes my favorite vibe benchmark too: “pelican riding a bicycle” 🪿🚲 - look at that beak 😄

000

Sujee Maniyam @sujee.dev · 13/01/2026

qwen3-next-80B-A3B-Thinking is now available on Nebius Token Factory. “Fast but verbose” per Artificial Analysis: artificialanalysis.ai/models/qwen3... My fav vibe benchmark: 'pelican riding a bicycle' 🪿🚲 Try it 👉 tokenfactory.nebius.com/playground?m... #OpenModels #LLM #PelicanRidingBicycle

000

Sujee Maniyam @sujee.dev · 08/01/2026

Installed @linuxmint.bsky.social on desktop & laptop Overall - smooth install - great collection of OS software Desktop: - detected dual monitors and NVIDIA GPU Laptop: - every thing works including suspend / wake A pic of my two fav mints :) Install notes: sujee.dev/post/another... #linux

020

Sujee Maniyam @sujee.dev · 21/11/2025

That's a wrap on two workshops at @qconferences.com Great crowd, good questions, and lots of fun discussions. 1️⃣ Chat with Your Website Using an LLM + Open Stack (Allycat) - Session: qconsf.com/training/nov... - Repo: github.com/The-AI-Allia...

121

Sujee Maniyam @sujee.dev · 19/11/2025

running 2 workshops Nov-20 at @qconferences.com 1. Chat with your website using an LLM and open stack (Allycat) qconsf.com/training/nov... 2. Open Source Rag Pipeline With Docling + Data Prep Kit + Milvus + Open LLMs qconsf.com/training/nov... discord: discord.gg/bk5fcvNJVZ #NebiusTokenFactory

011

Sujee Maniyam @sujee.dev · 17/10/2025

The AgentX–AgentBeats competition challenges you to "build agents that evaluate other agents" - how meta is that? 🤖 You’ll get to: 🌍 Build **public-good** systems through friendly competition 🏆 Compete for **$1M in prize money** 👉: rdi.berkeley.edu/agentx-agent... Proudly supported by #nebius

000

Sujee Maniyam @sujee.dev · 10/10/2025

Hah! Another workshop accepted at #QConSF 😃 "Allycat - chat with your website with LLMs" In this hands-on workshop, I will show how to crawl a website, index data and query it using natural language (not just keywords) with LLMs. qconsf.com/training/no... @qconferences.com #QConSF #docling

122

Sujee Maniyam @sujee.dev · 10/10/2025

Happy to be supporting NextBio Hackathon (AI x BIO Hackathon and DEMO Day) at at UCSF. luma.com/wuqccdje?tk... Connect, network and build together with other builders and AI x Biology development experts. Nebius AI Studio is sponsoring with credits. #NebiusAIStudio #Nebius

000

Sujee Maniyam @sujee.dev · 06/10/2025

I am running a workshop at QConSF on Nov 20, in SF. "Open Source Rag Pipeline With Docling + Data Prep Kit + Milvus + Open LLMs" You will walk away with working code you can build on. qconsf.com/training/no... #QConSF @qconferences.com #Milvus #RAG #DataPrepKit #Docling #NebiusAIStudio

101

Sujee Maniyam @sujee.dev · 01/10/2025

Code smarter with Qwen3 Coder in VSCode + Cline. Top open-source coding model, right in your editor. 🎥 Quick guide: www.youtube.com/watch?v=paK... #NebiusAIStudio #Nebius #Qwen3Coder #cline

010

Sujee Maniyam @sujee.dev · 30/09/2025

wondering about these model customizing techniques? Join me and Dylan Bristot for a webinar tomorrow Sept-30, 9am PST / 12 pm EST / 6pm CET 👉 nebius.com/events/webi... (even if you can't make it live, sign up, we will send you the recording and AI Studio credits!) #NebiusAI

000

Sujee Maniyam @sujee.dev · 29/09/2025

SVGENAI #4 meetup recap. - GPU accelerated storage - Jason Harmon - Using chatGPT to develop a chess playing Robot - Ryan O'Driscoll - Software engineering in the age of GenAI - Jules Damji of Databricks videos + slides : svgenai.dev/events/2025... Join us for future events: luma.com/svgenai

010

Sujee Maniyam @sujee.dev · 22/09/2025

Join us for great talks & networking 👉 luma.com/hmw007g7?ut... Sept 25 Thursday 530 pm Fremont #svgenai #meetup

000

Sujee Maniyam @sujee.dev · 16/09/2025

Distillation webinar - Sept 30 👉 nebius.com/events/webin... #distillation #LLM #webinar

000

Sujee Maniyam @sujee.dev · 27/08/2025

create awesome art using Nebius AI Studio guide: github.com/nebius/ai-st... quick video : www.youtube.com/watch?v=f060... #lora #nebiusai

010

Sujee Maniyam @sujee.dev · 19/08/2025

After the GPT-OSS deployment fix, Nebius now ranks as top-3 provider in Artificial Analysis benchmark on providers: artificialanalysis.ai/models/gpt-o... See before and after performance (images) Cause: mishandled 'reasoning-effort' flag! #nebius

000

Sujee Maniyam @sujee.dev · 09/08/2025

Friday fun - open source models wrap up - "pelican style" 🪿 This is my favorite 'vibe benchmark' comparing latest OS models (GPT-OSS, GLM 4.5, Qwen3-2507) released 😀 Prompt: create an svg of a pelican riding a bicycle Run all the latest open models on Nebius AI Studio - studio.nebius.com

010

Sujee Maniyam @sujee.dev · 18/07/2025

Docling is my go-to for document parsing. Used it in Allycat (github.com/The-AI-Allia...) + Data Prep Kit — handles PDFs, DOCX, HTML like a champ. Super easy to use. Just works. PDF benchmark: procycons.com/en/blogs/pdf... #Docling #OpenSource #AI #DataTools #PDFParsing @linuxfoundation.org

031

Sujee Maniyam @sujee.dev · 18/07/2025

😎 Cool project: make-it-heavy by Pietro Schirano. Dylan Bristot got it running on Nebius AI Studio. github.com/demianarc/ma... Tried it with QWEN3-30B-A3B 💥 Prompt: “Growth of solar power in US data centers” 4 agents → 1 solid report. Just so cool! 🧠 #MultiAgent #LLM #OpenSourceLLM

000

Sujee Maniyam @sujee.dev · 27/06/2025

Another #svgenai meetup is a wrap. 1) 📜 slides : svgenai.dev/meetups/2025... video coming soon. 2) 👉 Sign up: lu.ma/svgenai website: svgenai.dev 3) 💰 I gave out credits for Nebius AI Studio. studio.nebius.com. DM me if you'd like some. @davenielsen.bsky.social @ramnivas.com @melhq.bsky.social

010

Sujee Maniyam @sujee.dev · 26/06/2025

Devstral-Small-2505 - a top open-source #LLM optimized for coding agents, from @mistralai.bsky.social , available on Nebius AI Studio! ✨ #1 on SWE-Bench 📈 128K context window 🏛️ Apache 2.0 ✅ 🤗 huggingface.co/mistralai/De... Try it out : studio.nebius.com/playground?m... #devstral #nebiusAI

020

Sujee Maniyam @sujee.dev · 26/06/2025

🚀 AllyCat Update – Shiny New UI! - Sleek, modern interface - Toggle between dark and light themes - Built-in example prompts to help you get started - Transparent query processing – view retrieved documents from the vector DB 🎥: www.loom.com/share/360b09... github.com/The-AI-Allia... #AllyCat

000

Sujee Maniyam @sujee.dev · 25/06/2025

🚀 Talk: Building Multi-Agent Systems with Open Standards by @rodriveracom.bsky.social 🎁 Attendees get Nebius AI Studio credits! 📅 June 25 | ⏰ 10am PT / 1pm ET / 7pm CET 👉 Register: nebius.zoom.us/webinar/regi... #MultiAgent #OpenStandards

020

Sujee Maniyam @sujee.dev · 13/06/2025

LiteLLM now supports all models on Nebius AI Studio. Run all the latest open source models (DeepSeek-R1-0528, Qwen3-235B-A22B) and many more : studio.nebius.com doc & examples : docs.litellm.ai/docs/provide... code : github.com/nebius/ai-st... #LiteLLM #NebiusAIStudio

000

Sujee Maniyam @sujee.dev · 26/03/2025

my upcoming talk: Create High-Quality Datasets by Filtering Out Spam, HAP (Hate, Abuse, Profanity) Speech, and Sensitive Data 🗓️: Thursday Mar 27, 2025 ⏰: 9am PST / 12pm EST Register: lnkd.in/gfRQyzKZ #dataprepkit #LLM #AIAlliance

000

Sujee Maniyam @sujee.dev · 10/03/2025

Practical tips for running an effective unconference sujee.dev/post/unconfe... #tips #unconference

000

Sujee Maniyam @sujee.dev · 06/03/2025

Talk: How was GneissWeb ("niceWeb") - a state-of-the-art LLM pre-training dataset with ~10 Trillion tokens - created 🤗: huggingface.co/datasets/ibm... 🗓️: Mar 06, '25 ⏰: 9 am PST / 12 pm EST / 5pm GMT Signup: tinyurl.com/yydfeyv7 @aialliance.bsky.social @ibm-research.bsky.social #LLM #opendata

010