1. X
  2. StepFun
Log inSign up
StepFun
376 posts
Image
user avatar
StepFun
@StepFun_ai
Scale-up possibilities for everyone. OpenRouter: openrouter.ai/stepfun HuggingFace: huggingface.co/stepfun-ai
stepfun.com
Joined February 2025
157
Following
11.5K
Followers
AffiliatesAffiliatesRepliesRepliesMediaMedia

New to X?

Sign up now to get your own personalized timeline!

Create account

By signing up, you agree to the Terms of Service and Privacy Policy, including Cookie Use.

Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
Don't miss what's happening
People on X are the first to know.
Log inSign up
  • Pinned
    user avatar
    StepFun
    @StepFun_ai
    May 29
    ⚡️ Step 3.7 Flash is here: The new frontier is agent efficiency. #1 ClawEval-1.1 (67.1), #1 SimpleVQA Search (79.2), #2 SWE-PRO (56.3), 95.3 on V* Python. Open weights under Apache 2.0. Built for agentic, coding, search, and multimodal workflows — balancing speed, cost, and
    Image
    350K0350K
  • user avatar
    StepFun
    @StepFun_ai
    Jul 16
    🟠 StepFun is at #WAIC2026 #StepFun 🚗 ZEEKR 8X × Super EvaStep into StepFun Amoo World — a treasure-hunt map across 7 stops: 🤖 robots building the Great Wall 🎮 arcade penalty shootout w/ Amoo 📱 STEPX agent phone live demos 🗓️ Jul 17–20 📍 Shanghai World Expo Center · Hall
    Image
    Made with AI
    1.5K01.5K
  • user avatar
    StepFun
    @StepFun_ai
    Jul 15
    Great to see Step 3.7 Flash now available through @baseten. We built this model for real-world agent workflows, with native multimodal understanding, reliable tool use and the efficiency needed for production. Thanks to the Baseten team for the great collaboration.
    user avatar
    Baseten
    @baseten
    Jul 14
    Step 3.7 Flash is now in the Baseten Model Library! This is a 198B-parameter sparse MoE model (11B active per token) with native image and video input, and a 256K context window. It's a strong option for visual reasoning, agentic coding, and long-context tasks.
    Image
    5.4K05.4K
  • user avatar
    StepFun
    @StepFun_ai
    Jun 30
    Step 3.7 Flash is currently top 10 on @OpenRouter this month, with 4.29T tokens routed. Builders are pushing it through real agent runs, coding tasks, and long-context workflows. Keep sending the hard stuff.
    Image
    8.6K08.6K
  • user avatar
    StepFun
    @StepFun_ai
    Jun 30
    Been great seeing Step 3.7 Flash get real use in Nous Portal: people testing, building, and running all kinds of agent workflows with it. We’re keeping free access going with the @NousResearch! Try it out and send us what you make.
    user avatar
    Nous Research
    @NousResearch
    Jun 30
    More time to build with Step 3.7 Flash: in partnership with @StepFun_ai, we’re extending the free usage period in Nous Portal by an additional 15 days!
    Image
    00:00
    5.6K05.6K
  • user avatar
    StepFun
    @StepFun_ai
    Jun 29
    Step 3.7 Flash hits #2 on Claw-Eval General for autonomous agents. We’re seeing strong performance across multi-step execution and robustness in long-horizon tasks, ranking just behind Claude Opus 4.6. Promising signals for real-world agent workloads.
    Image
    25K025K
    user avatar
    StepFun
    @StepFun_ai
    Jun 29
    Claw-eval Link: claw-eval.github.io
    1.8K01.8K
  • user avatar
    StepFun
    @StepFun_ai
    Jun 24
    This is the kind of agent workflow Step Plan was built for: connect it once, push through a real build, and keep experimenting without thinking about every single API call. Love the tarot generator demo. Thanks for testing Step 3.7 Flash in Claude Code, @codedailyML 🙌
    user avatar
    Z-Coder
    @codedailyML
    Jun 22
    I used to dread heavy testing days because every API call felt like watching money disappear in real time. Found a backend built to run flat instead of per call, made specifically for agent work. Pointed my Claude Code setup at it and had it running in under 5 minutes. Here's
    Image
    00:00
    4.9K04.9K
  • user avatar
    StepFun
    @StepFun_ai
    Jun 24
    This is the pain we kept hearing from builders: once an agent starts doing real work, the meter becomes part of the workflow. Step Plan is our attempt to make that less of a distraction. Thanks for putting Step Plan + Step 3.7 Flash through a real Claude Code setup 🙌
    user avatar
    Chidanand Tripathi
    @thetripathi58
    Jun 22
    I almost stopped testing new models altogether. Not because they were bad. Because every call left a number climbing in the corner of my screen, and at some point I started watching that number more than I was building. A flat-rate backend fixed that for agent work. I swapped
    Image
    00:00
    3.8K03.8K
  • user avatar
    StepFun
    @StepFun_ai
    Jun 22
    Introducing the StepFun Startup Program. We’re supporting early-stage AI teams building real products with StepFun models — from multimodal applications to agentic systems. Selected startups may receive API credits, dedicated ecosystem support, co-marketing opportunities,
    Image
    17K017K
  • StepFun reposted
    user avatar
    Franky.
    StepFun
    @FrankYouChill
    Jun 18
    Tried Step 3.7 Flash in @cline with Playwright MCP. - Select the Step 3.7 Flash model (free) - Installed Playwright MCP through one-click prompt - Let it open the web page, interact with it, and capture screenshots This shows the workflow: efficient agentic model + good tools +
    Image
    00:00
    4.3K04.3K
  • user avatar
    StepFun
    @StepFun_ai
    Jun 18
    Excited to partner with @cline to bring Step 3.7 Flash to more coding agent builders. Step 3.7 Flash is now free in Cline for the next month — capable, fast, and reliable for agentic coding workflows. Try it in Cline: /model → Step 3.7 Flash Big thanks to the Cline team.
    user avatar
    Cline
    @cline
    Jun 17
    Step 3.7 Flash is free in Cline for the next month. It beats Gemini and DeepSeek flash models, and comes surprisingly close to frontier performance on SWE Bench. Open weights, 256k context window, fast and reliable. npm i -g cline Run cline > use /model > select Step 3.7 Flash
    Image
    3.1K03.1K
  • StepFun reposted
    user avatar
    Md Riyazuddin
    @riyazmd774
    Jun 11
    One prompt. One model. A complete fitness app in seconds. 🔥 I tested the new StepFun Step 3.7 Flash and gave it this prompt: "Build a beautiful single HTML fitness tracker app using Tailwind CSS, Chart.js, and Font Awesome. Dark modern theme. 4 tabs Dashboard, Workouts, Diet,
    Image
    00:00
    Paid partnership
    23K023K
  • user avatar
    StepFun
    @StepFun_ai
    Jun 16
    Excited to see Step 3.7 Flash live via @novita_labs on @OpenRouter. Built for high-efficiency agent workloads, Step 3.7 Flash combines native multimodal understanding, strong agentic coding capabilities, reliable tool use, and web & visual search workflows for production AI
    user avatar
    Novita AI
    @novita_labs
    Jun 16
    🚀 Step 3.7 Flash from @StepFun_ai is now available via Novita on @OpenRouter. High-efficiency multimodal reasoning. Multi-step tool use. Built for coding and agent workflows.
    Image
    7.1K07.1K
  • user avatar
    StepFun
    @StepFun_ai
    Jun 14
    Step 3.7 Flash is now live on @DeepInfra 🚀 Builders and teams can now try our open-source multimodal reasoning model through DeepInfra’s API, with private endpoint deployment available for dedicated workloads. Built for agentic coding, tool use, search, and vision workflows.
    user avatar
    DeepInfra
    @DeepInfra
    Jun 13
    Step 3.7 Flash is Live on DeepInfra: An Agentic, Multimodal Model Built for Production
    Image
    11K011K
Advertisement
Advertisement