Kaige

Making inference more affordable and keeping your data private.

[Xiaomi Tech @XiaomiTech_]

Real tasks test real capabilities.

Three weeks after launch, Artificial Analysis and Arena released four real-world results for Xiaomi MiMo-V2.6:

• FrontierCode 1.1 (real-repo coding): Pro ranks #1 open-source, entering the capability–cost Pareto frontier
• Cyber Index (vulnerability defense): Pro ranks #1 open-source, entering the Pareto frontier
• Agent Arena (real user sessions): Pro ranks #1 open-source; both Pro & Flash enter the Pareto frontier
• AA-3D (interactive 3D): Pro ranks #2 open-source

#XiaomiAI #MiMo #XiaomiMiMo

[Kaige @Kiango20]

Text, image & video APIs on Inferway.
MiMo-V2.6 Flash · Z-Image Turbo · MiniMax H3.

Free to use within daily limits, affordable when you need more.
Text prompts and replies aren't stored.
OpenAI & Anthropic-compatible text APIs.

https://t.co/FAc3pNdN0t

[Serhii Sternenko @sternenko]

❗️Last ​​night, the operation of Starlink terminals in Ukraine was restricted.
Connectivity began to drop when moving at speeds exceeding 160 km/h. As a result, a number of critical missions were disrupted.

SpaceX likely took these measures at the request of the Trump administration. Also last night, when Kyiv was attacked by ballistic missiles, Ukraine did not receive intelligence regarding their launch— a failure that could have led to higher casualties.

Ukraine did not attack Russian oil refineries, yet Trump has already begun imposing sanctions on my country. Meanwhile, sanctions against Russia have been eased.

[Kaige @Kiango20]

https://t.co/bh342Bu86J turned one month old this week 🎉

To say thanks, we just added free credit to every active account: $10 (or ¥20 for China accounts). No expiry, no strings. It's already in your balance.

Thanks for building with us.

[Kaige @Kiango20]

@soy_muse https://t.co/FAc3pNdN0t , welcome to have a try!

[Kaige @Kiango20]

MiMo-V2.6 Flash goes beyond chat. Give it a tricky problem, some code, or an image to analyze—it can use tools, too.

MiniMax H3 768P turns text into video.

MiMo prompts and replies aren’t stored; H3 clips stay for 7 days.
Both have limited free use.

https://t.co/FAc3pNekQ1

[Kaige @Kiango20]

MiMo-V2.6 Flash for chat. MiniMax H3 768P for video.

You can try both on Inferway. MiMo handles chat without storing prompts or replies; H3 turns text into video, with clips kept for 7 days.

Both have limited free options

https://t.co/FAc3pNdN0t

[Tech2Wild @Tech2Wild]

Mimo V2.6 Flash Comes in behind Deepseek V4.1 Flash !

[Kaige @Kiango20]

Two models, two ways to create on Inferway:

⚡ MiMo-V2.6 Flash — chat, brainstorm, and use tools.
🎬 MiniMax H3 768P — turn a text prompt into video with audio.

Free access is limited; H3 requires sign-in.

https://t.co/FAc3pNekQ1

[Kaige @Kiango20]

@ilovek8s https://t.co/FAc3pNdN0t 这里H3 用起来就像不要钱一样

[Kaige @Kiango20]

MiMo-v2.6-Flash is now live on https://t.co/FAc3pNdN0t! ⚡️

Practically free—use it like a billionaire without spending like one. Go try it out now!

[Xiaomi MiMo @XiaomiMiMo]

Introducing Xiaomi MiMo-V2.6 — Pro & Flash.

Frontier intelligence, all the modalities, built in public.

🔹 Two omnimodal models, advancing through scaled reinforcement learning
🔹 Pro performs on par with Claude Opus 5 and GPT-5.6 Sol across most agent benchmarks
🔹 Pro scores 46 on the Artificial Analysis Intelligence Index — the highest among open-source models
🔹 Stronger coding, computer use, 3D reasoning and creative capabilities
🔹 Open model weights, technical report, RL environments and training code

Blog：https://t.co/JGdwC9Ymvj

[Kaige @Kiango20]

MiMo-v2.6-Flash is officially on https://t.co/FAc3pNekQ1 🚀

Top-tier intelligence at practically zero cost. Run inference like a billionaire—try it free today!

[Xiaomi MiMo @XiaomiMiMo]

Introducing Xiaomi MiMo-V2.6 — Pro & Flash.

Frontier intelligence, all the modalities, built in public.

🔹 Two omnimodal models, advancing through scaled reinforcement learning
🔹 Pro performs on par with Claude Opus 5 and GPT-5.6 Sol across most agent benchmarks
🔹 Pro scores 46 on the Artificial Analysis Intelligence Index — the highest among open-source models
🔹 Stronger coding, computer use, 3D reasoning and creative capabilities
🔹 Open model weights, technical report, RL environments and training code

Blog：https://t.co/JGdwC9Ymvj

[Inferway @Inferway_Ai]

Two models,One platform. Zero friction. ⚡️

Now live on Inferway:
🧠 MiMo-V2.6 Flash (Text + Vision) | 262K context & tool calling
🎬 MiniMax H3 768P (Video + Audio) | 768P dynamic clips (5–10s)
🔥 UP TO 70% OFF inference costs.

[Inferway @Inferway_Ai]

Minimax H3 is now live on Inferway!🚀

Experience faster, safer, and more cost-effective AI inference today!🤩🤩

Try it now ⬇️
https://t.co/FtEPIDkYWR

[Mark Jaquith @markjaquith]

This is my new shortest “what is Jev” explanation

[Kaige @Kiango20]

@supezen https://t.co/FAc3pNdN0t 做了个self-host模型站, 陆陆续续上模型ing. 一共用掉了几千亿token, 还没算本地模型的用量