Freedom to reliably deploy any open model effortlessly. Unlimited tokens for 30K+ models. Open-source jev models on https://xitter.cf/t.co/HVsStpj1i3
San Francisco
Joined June 2024
- Tweets586
- Following129
- Followers1.9K
- Likes683
Pinned Tweet
Today we're announcing our $20 M Series A, co-led by @AMDVentures and @AirbusVentures, with @BMWiVentures, @Kickstartph, @PanacheVC, and @Wavemaker_VC.
We're building the neutral layer for open models — so the future doesn't run on a few labs, a few chips, or a few clouds.
THE FUTURE CAN’T BE BLOCKED.
Any Model. Anywhere. Anyone. Yours.
We’ve seen people still running Llama 2 in 2026.
Too many teams have no idea what models they’re actually running, what they’re paying for them, or why they chose them.
That’s why our CEO @picocreator took the stage at @aiDotEngineer NYC with a workshop on exploding token bills, and how to get yours back under control.
Full talk below!
youtube.com/watch?v=Ti-BmGU1…
Space Bunny is here!
Congrats to @MiniMax_AI on the launch of MiniMax M3.1 Flash 🚀
It’s the best open-weights model on @ArtificialAnlys’ AutomationBench, and it costs a fraction of what closed models charge per task.
Open models keep getting better and cheaper.
The weights drop on Oct 23, live on Featherless the same day:
featherless.ai/model-pages/m…
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: platform.minimax.io
M Plan: minimax.io/m-plan
MiniMax Code for desktop: agent.minimax.io/download
Thank you again to @StepFun_ai for making this possible.
xitter.cf/stepfun_ai/status/2110…
Step 5 Preview is now open weights.
BF16 and FP8 weights are live on Hugging Face and ModelScope, together with our technical blog.
• Sparse MoE: 600B total, 27B active per token
• 1M-token context window
• Native text and vision input
• Built for long-horizon agentic work: coding, software engineering, professional knowledge work and finance
Frontier intelligence, open weights. Build what's next.
Hugging Face: huggingface.co/stepfun-ai/St…
ModelScope: modelscope.ai/models/stepfun…
GitHub: github.com/stepfun-ai/Step-5…
Under the hood:
- 92-layer narrow-and-deep architecture for implicit multi-hop reasoning without blowing up compute
- Sparse GQA with 8-token block indexing, cutting indexing and top-k overhead to 1/8 at million-token context
- FP32 for key states plus fine-grained gradient clipping to keep deep-network training stable
- RL on million-scale verifiable environments, with context compression built into end-to-end policy training
Live on Featherless!
What it is good at:
- Fixing bugs and shipping features across unfamiliar codebases, with changes that run and pass checks
- Turning screenshots, mockups, or PRDs into working web pages, or building browser games from assets
- Finance analysis from annual reports and operating data, delivered as reports, spreadsheets, and decks
Step 5 Preview is here.
@StepFun_ai's new open-weights model sets a new Pareto frontier, matching Kimi K3 on the @ArtificialAnlys Intelligence Index!
It’s built for long-horizon agent work, with frontier-level results on software engineering and professional knowledge work, especially finance.
- 600B sparse MoE, 27B active per token
- ~1/6 the cost per task of Claude Opus 5.5, per Artificial Analysis
Now on Featherless, served at FP8.
featherless.ai/model-pages/s…
Featherless AI retweeted
Free place to stay in West NYC during @aiDotEngineer
Oct 12–14: a private room
Oct 14–18: a whole 3-bedroom apartment to yourself
If you’re building in AI, we at @featherlessai would love to host you. DM is open.
The Featherless team is LIVE at @aiDotEngineer NYC!
Join us as @picocreator delivers a talk on Tokenomics -- how to grab back your out-of-control token spending!
Today in AI:
Can frontier language models control a robot?
The answer is, to some extent, yes!
None of these models were trained to drive a robot arm. They're writing "hello world" with a pen, zero-shot.
Our prediction: the next major upcoming frontier for open source AI is robotics, it is simply just bottlenecked by the physical world.
Robots need models that run locally, cheaply, and that you can fine-tune on your own data, which is where open source models fit perfectly.
Today, we are announcing Featherless.
It is here: featherless.ai/
Today in Open Source AI:
Step 5 Preview!
- 600B parameters total
- Only 27B active per token, about 4.5% of the model being activated any time
- 1M tokens in and out
- About 1/8 the cost per task of Claude Opus 5 (per Artificial Analysis)
Check it out 👀
Step 5 Preview now exists where you build.
Now live on @OpenRouter. A week of free access is rolling out today across @opencode, @cline , @NousResearch, @kilocode, and more.
Flagship-tier intelligence for agentic and professional work, at substantially lower task cost.
Keep your workflow. Switch your model.
Try Step 5 Preview on OpenRouter: openrouter.ai/stepfun/step-5…
Featherless AI retweeted
Introducing Feather Pop, playable in your browser.
I have been having a blast creating retro style games lately.
Thank you to @StepFun_ai for giving me access to Step 5 Preview, it wrote the vast majority of the code and did a great job.
We're entering a new era of entertainment, and the models will only continue to get bettter.
Open weights are coming Oct 15 to Featherless, but the game is out now, check out the links below!
Featherless AI retweeted
The SF Tech Week filter is now live, and free.
Some events are fake: either they don't exist, or they're a 1:1 sales pitch in disguise.
1,697 events. Filter them all with SimpleJev 👇
