Video yükleniyor...

Video Yüklenemedi

Ana Sayfaya Dön

MiniMax M2.1 is officially live🚀 Built for real-world coding and AI-native organizations — from vibe builds to serious workflows. A SOTA 10B-activated OSS coding & agent model, scoring 72.5% on SWE-multilingual and 88.6% on our newly open-sourced VIBE-bench, exceeding leading closed-source models like Gemini 3 Pro and Claude 4.5...

MiniMax_AI's profile picture

MiniMax (official)

130,848 subscribers

1,074,832 görüntüleme • 10 ay önce •via X (Twitter)

EğitimHaberler & PolitikaBilim & Teknoloji
LIVE

Anya Rossi• Live Now

Private livecam show

32 Yorum

MiniMax (official) profil fotoğrafı
MiniMax (official)10 ay önce

SOTA across SWE-Verified, SWE-Multilingual, Multi-SWE, VIBE-Bench, and Terminal-Bench 2.0.

MiniMax (official) profil fotoğrafı
MiniMax (official)10 ay önce

API now available on MiniMax Open Platform: MiniMax Agent (powered by M2.1): MiniMax Developer Discord Read more on Full open-source release in 2 days, stay tuned

MiniMax (official) profil fotoğrafı
MiniMax (official)10 ay önce

We would like to thank our launch partners @Ollama @FactoryAI @cline @kilocode @roocode @blackboxai @openrouter @vercel @opencode and developers/ambassadors for your early testing feedback and support!

Papersnake profil fotoğrafı
Papersnake10 ay önce

A little better than GLM-4.7 on SWE

ZOYA ✪ profil fotoğrafı
ZOYA ✪10 ay önce

Open-sourcing at this level while outperforming closed models is impressive. Strong signal for where agentic OSS is headed.

lord asado 🥩 profil fotoğrafı
lord asado 🥩10 ay önce

It’s raining language models for Christmas.

Robin Ebers profil fotoğrafı
Robin Ebers10 ay önce

why no opus 4.5 thinking comparison go big or go home imo

Eric Hartford profil fotoğrafı
Eric Hartford10 ay önce

I can't wait to see it on @huggingface !

Anes Valentic profil fotoğrafı
Anes Valentic10 ay önce

What happened with the weights? I’m unable to find them…

KwaiKAT profil fotoğrafı
KwaiKAT10 ay önce

Great work on M2.1! 🚀 Impressive benchmarks — exciting to see M2.1 pushing OSS coding models forward.

Arsalan profil fotoğrafı
Arsalan10 ay önce

the fact that it beats Claude 4.5 and Gemini 3 Pro at this size? Huge for open-source devs.

Hamza profil fotoğrafı
Hamza10 ay önce

Really impressed by the quality of the front-end design, which was built from scratch in one shot This is excellent progress, congratulations to the team

Manish Kulariya profil fotoğrafı
Manish Kulariya10 ay önce

Looks like a big week for open-source models: yesterday GLM, and now MiniMax, with almost the same SWE-Bench numbers. Great work!

goosewin profil fotoğrafı
goosewin10 ay önce

Christmas is here early!

Mr. Ånand profil fotoğrafı
Mr. Ånand10 ay önce

Time to use it within agent system👀

Fahd Mirza profil fotoğrafı
Fahd Mirza10 ay önce

🚀 MiniMax-M2.1 just dropped and it's INSANE ✨ What's new: CRUSHES multi-language coding (72.5% on SWE-bench) Beats Gemini 3 Pro & Claude Sonnet 4.5 8 languages: Rust, Java, Go, C++, Kotlin, Obj-C, TS, JS Mobile & web dev? ELITE Lightning fast ⚡ Less tokens = lower costs 💰 🔥 Open source in 2 DAYS 💪Watch video here:

ZaxtR profil fotoğrafı
ZaxtR10 ay önce

On the brand-new open-sourced VIBE-Bench (Visual & Interactive Benchmark for Execution)—a rigorous full-stack app development eval covering web, Android, iOS, simulation, and backend—M2.1 scores an impressive 88.6% overall (with highs like 91.5% on VIBE-Web and 89.7% on VIBE-Android), surpassing closed-source leaders like Gemini 3 Pro and Claude 4.5 Sonnet in several categories.

Contextrix profil fotoğrafı
Contextrix10 ay önce

72.5% on SWE-multilingual is massive for an open-weight model with only 10B active params. Beating closed giants like Claude 4.5 Sonnet on VIBE-bench proves efficient models are catching up fast for real coding work.

meron the melon profil fotoğrafı
meron the melon10 ay önce

soooooo COOL

Pietro Montaldo profil fotoğrafı
Pietro Montaldo10 ay önce

MiniMax M2.1 is next-level OSS

Poonam Soni profil fotoğrafı
Poonam Soni10 ay önce

Open source is winning

Emily profil fotoğrafı
Emily10 ay önce

Open source 🔥🔥🔥

Hassan profil fotoğrafı
Hassan10 ay önce

Congrats on the launch! Excited to try this out

tobitege profil fotoğrafı
tobitege10 ay önce

What about C#?

petebray profil fotoğrafı
petebray10 ay önce

Get this into Cursor please. W/o having to dance w API keys.

zoomies.llamafile profil fotoğrafı
zoomies.llamafile10 ay önce

@GroqInc needs this

Len Seaside profil fotoğrafı
Len Seaside10 ay önce

10B!? That's incredible. Will be on huggingface soon? Brilliant work guys.

combif1am profil fotoğrafı
combif1am10 ay önce

Hi

Joaquin Bonifacino profil fotoğrafı
Joaquin Bonifacino10 ay önce

A 10B model better than sonnet 4.5 and gpt 5.2? What?

AmenoTurbo (∎, ∆) profil fotoğrafı
AmenoTurbo (∎, ∆)10 ay önce

Congratulations ! I tried to run minimaxM2 locally on my strix halo machine, due to issues with vllm I use llama.cpp but I don't manage to make interleaved thinking work. What could I do to make interleaved thinking work on llamap.cpp + openwebui Answers would be much appreciated

Ersin KOÇ profil fotoğrafı
Ersin KOÇ10 ay önce

FluidFlow & Minimax M2.1

Vignesh G profil fotoğrafı
Vignesh G10 ay önce

@UnslothAI

Benzer Videolar

Nothing beats open-source! MiniMax just dropped M2.1, and devs are calling it "Claude at 10% the cost." - 72.5% SWE-Multilingual. Beats Sonnet 4.5 - 88.6% VIBE-bench. Beats Gemini 3 Pro I used it to build an AI studio that turns any website into a podcast. 100% open-source.
2:47

Nothing beats open-source! MiniMax just dropped M2.1, and devs are calling it "Claude at 10% the cost." - 72.5% SWE-Multilingual. Beats Sonnet 4.5 - 88.6% VIBE-bench. Beats Gemini 3 Pro I used it to build an AI studio that turns any website into a podcast. 100% open-source.

Akshay 🚀

140,572 görüntüleme • 10 ay önce

Introducing Gemini 3 - Google's most powerful model for web development, "vibe coding" and agentic tasks.
2:36

Introducing Gemini 3 - Google's most powerful model for web development, "vibe coding" and agentic tasks.

Addy Osmani

43,677 görüntüleme • 11 ay önce

MiniMax M2.1 is really good for writing code. It's open source, and you can connect Visual Studio Code to it and start coding right away. The model is especially good for those who don't use Python as a primary language. It's currently the best multilingual model out there. The model hits 72.5% on SWE-Multilingual and 88.6% on VIBE-Bench (best!) This means the model is also really good at performing real engineering work and does very well at agentic tasks: • Long tool chains • Structured agent scaffolding • Following repo-level rules and slash commands • Working inside existing agent frameworks without falling apart And obviously, it's faster, cheaper, and more concise than M2, which was already really good. I recorded a video to show you how to run MiniMax (official)'s 2.1 using Visual Studio Code.
5:45

MiniMax M2.1 is really good for writing code. It's open source, and you can connect Visual Studio Code to it and start coding right away. The model is especially good for those who don't use Python as a primary language. It's currently the best multilingual model out there. The model hits 72.5% on SWE-Multilingual and 88.6% on VIBE-Bench (best!) This means the model is also really good at performing real engineering work and does very well at agentic tasks: • Long tool chains • Structured agent scaffolding • Following repo-level rules and slash commands • Working inside existing agent frameworks without falling apart And obviously, it's faster, cheaper, and more concise than M2, which was already really good. I recorded a video to show you how to run MiniMax (official)'s 2.1 using Visual Studio Code.

Santiago

47,879 görüntüleme • 9 ay önce

📣 Claude Sonnet 4.5, Anthropic’s most advanced model for coding and real-world agents, is now rolling out in GitHub Copilot to Copilot Pro, Pro+, Business, and Enterprise ⬇️
1:53

📣 Claude Sonnet 4.5, Anthropic’s most advanced model for coding and real-world agents, is now rolling out in GitHub Copilot to Copilot Pro, Pro+, Business, and Enterprise ⬇️

GitHub

179,854 görüntüleme • 1 yıl önce

Today we’re taking a big step on the path toward AGI and releasing Gemini 3— our most intelligent model yet. With Gemini 3, you can bring any idea to life. It is state-of-the-art in reasoning, the best model in the world for multimodal understanding, and our best agentic and vibe coding model.
0:51

Today we’re taking a big step on the path toward AGI and releasing Gemini 3— our most intelligent model yet. With Gemini 3, you can bring any idea to life. It is state-of-the-art in reasoning, the best model in the world for multimodal understanding, and our best agentic and vibe coding model.

Google AI

493,668 görüntüleme • 11 ay önce

GPT-5 is live across the Vercel AI Cloud, and it's available: • on AI Gateway • as a model in • in our new open-source vibe coding platform
1:24

GPT-5 is live across the Vercel AI Cloud, and it's available: • on AI Gateway • as a model in • in our new open-source vibe coding platform

Vercel

15,018 görüntüleme • 1 yıl önce

Start building with Gemini 3 Pro, our most intelligent model with state-of-the-art reasoning and complex multimodal understanding, as well as powerful agentic and vibe coding capabilities. See more below:
0:20

Start building with Gemini 3 Pro, our most intelligent model with state-of-the-art reasoning and complex multimodal understanding, as well as powerful agentic and vibe coding capabilities. See more below:

Google AI Developers

104,898 görüntüleme • 11 ay önce

The MiniMax M2 model is mind-blowing! It's open-source. It outperforms Gemini 2.5, Claude 4.1, and Qwen3 across coding and tool-use benchmarks. Right now, it's one of the world's top 5 models in intelligence! And here is the best part: Claude is one of the best models you can use today, and MiniMax M2 costs only 8% of that! It's smaller, faster, and cheaper. Extremely efficient at using tokens. Minimax M2's biggest strength: High agentic capabilities. The model can plan and execute complex multi-tool workflows. It's reliable and very robust at executing long-horizon tool chains. In summary: • Low latency • Very cheap • Excels at agentic tasks • Open-source The model currently powers the MiniMax Agent and is available for a free global trial. You can access MiniMax M2's API here: To access the agent: And here is the MiniMax website: Thanks to the MiniMax team for showing me the ropes and partnering with me on this post.
6:01

The MiniMax M2 model is mind-blowing! It's open-source. It outperforms Gemini 2.5, Claude 4.1, and Qwen3 across coding and tool-use benchmarks. Right now, it's one of the world's top 5 models in intelligence! And here is the best part: Claude is one of the best models you can use today, and MiniMax M2 costs only 8% of that! It's smaller, faster, and cheaper. Extremely efficient at using tokens. Minimax M2's biggest strength: High agentic capabilities. The model can plan and execute complex multi-tool workflows. It's reliable and very robust at executing long-horizon tool chains. In summary: • Low latency • Very cheap • Excels at agentic tasks • Open-source The model currently powers the MiniMax Agent and is available for a free global trial. You can access MiniMax M2's API here: To access the agent: And here is the MiniMax website: Thanks to the MiniMax team for showing me the ropes and partnering with me on this post.

Santiago

91,300 görüntüleme • 1 yıl önce

Introducing Gemini 3 ✨ It’s the best model in the world for multimodal understanding, and our most powerful agentic + vibe coding model yet. Gemini 3 can bring any idea to life, quickly grasping context and intent so you can get what you need with less prompting. Find Gemini 3 Pro rolling out today in the Google Gemini and AI Mode in Search. For developers, build with it now in Google AI Studio and Vertex AI. Excited for you to try it!
0:51

Introducing Gemini 3 ✨ It’s the best model in the world for multimodal understanding, and our most powerful agentic + vibe coding model yet. Gemini 3 can bring any idea to life, quickly grasping context and intent so you can get what you need with less prompting. Find Gemini 3 Pro rolling out today in the Google Gemini and AI Mode in Search. For developers, build with it now in Google AI Studio and Vertex AI. Excited for you to try it!

Sundar Pichai

2,956,409 görüntüleme • 11 ay önce

Starting today, Gemini 3 is now available for developers. Built to accelerate software development, it can help you bring your ideas to life no matter your level of experience. Our first model in the series, Gemini 3 Pro, is incredibly strong at: ✨ Agentic coding ✨ Vibe coding and web dev ✨ Multimodal understanding
0:41

Starting today, Gemini 3 is now available for developers. Built to accelerate software development, it can help you bring your ideas to life no matter your level of experience. Our first model in the series, Gemini 3 Pro, is incredibly strong at: ✨ Agentic coding ✨ Vibe coding and web dev ✨ Multimodal understanding

Google

93,051 görüntüleme • 11 ay önce

🔗 Announcing LangChain OSS Skills LangChain has the most popular frameworks for building AI agents — and now your coding agent can be an expert in it. We're excited to release the first iteration of LangChain OSS Skills, giving your agent expertise in our open source frameworks. The skills include guidance on how to use langchain, langgraph, and deepagents to effectively build agents. ➡️ Install our OSS skills for your coding agent here: ➡️ Read more:
0:43

🔗 Announcing LangChain OSS Skills LangChain has the most popular frameworks for building AI agents — and now your coding agent can be an expert in it. We're excited to release the first iteration of LangChain OSS Skills, giving your agent expertise in our open source frameworks. The skills include guidance on how to use langchain, langgraph, and deepagents to effectively build agents. ➡️ Install our OSS skills for your coding agent here: ➡️ Read more:

LangChain OSS

65,840 görüntüleme • 8 ay önce

Minimax M3 is excellent at SVG generation, reaching close to Gemini 3.5 Flash levels and beating Opus 4.7 on SVG-Bench. With 1M context, native multimodality, strong agentic/coding ability and open weights coming soon, the closed-source moat is thinning fast. Full Video:
0:22

Minimax M3 is excellent at SVG generation, reaching close to Gemini 3.5 Flash levels and beating Opus 4.7 on SVG-Bench. With 1M context, native multimodality, strong agentic/coding ability and open weights coming soon, the closed-source moat is thinning fast. Full Video:

WorldofAI

16,499 görüntüleme • 5 ay önce

Check out the OSS Coding Agent template, powered by AI SDK, Vercel AI Gateway and Sandbox. We now have open source, BYO-model versions of both v0 and Vercel Agent 😲
0:25

Check out the OSS Coding Agent template, powered by AI SDK, Vercel AI Gateway and Sandbox. We now have open source, BYO-model versions of both v0 and Vercel Agent 😲

Guillermo Rauch

65,851 görüntüleme • 1 yıl önce

Spent the last few days vibe coding on my NVIDIA DGX Spark. Here's what I learned. Qwen 3.5 122B took one minute and nine seconds to respond "Hi how are you doing". Unusable for vibe coding. Gemma 4 was fast but built a dot instead of a first person shooter game. GPT-OSS 120B was the sweet spot. Fast, capable, and actually produced working HTML. Open source models running locally are not replacing Claude Opus 4.6 or Codex with GPT 5.4. Not even close. But they're getting better every month. The new DGX Spark Bench is live on Real-world benchmarks for local models on local hardware. This is just the start. Full video below.
21:03

Spent the last few days vibe coding on my NVIDIA DGX Spark. Here's what I learned. Qwen 3.5 122B took one minute and nine seconds to respond "Hi how are you doing". Unusable for vibe coding. Gemma 4 was fast but built a dot instead of a first person shooter game. GPT-OSS 120B was the sweet spot. Fast, capable, and actually produced working HTML. Open source models running locally are not replacing Claude Opus 4.6 or Codex with GPT 5.4. Not even close. But they're getting better every month. The new DGX Spark Bench is live on Real-world benchmarks for local models on local hardware. This is just the start. Full video below.

BridgeMind

57,304 görüntüleme • 7 ay önce

Here’s how to build a Slack clone with Claude 3.7 Sonnet and Cursor Agent in v0.46. Perfect for *anyone* wanting to pick up AI coding tools. Learn how to build an app 100% from scratch with the newest AI coding workflows. Watch for 1hr+ of agentic coding.
1:12:45

Here’s how to build a Slack clone with Claude 3.7 Sonnet and Cursor Agent in v0.46. Perfect for *anyone* wanting to pick up AI coding tools. Learn how to build an app 100% from scratch with the newest AI coding workflows. Watch for 1hr+ of agentic coding.

Mckay Wrigley

212,638 görüntüleme • 1 yıl önce

In <4 months, Ridges used an open, incentivized tournament to produce the top open-source coding agent. • It outperforms all open-source agentic systems on SWE-Bench • Beats closed models from billion-dollar labs • Was done with $0 VC funding
21:31

In <4 months, Ridges used an open, incentivized tournament to produce the top open-source coding agent. • It outperforms all open-source agentic systems on SWE-Bench • Beats closed models from billion-dollar labs • Was done with $0 VC funding

Sami Kassab

58,074 görüntüleme • 1 yıl önce

Open-ended coding training data may no longer be the bottleneck: AI can scale open-ended tasks—and even outperform human-expert curation. FrontierCS team is releasing FrontierSmith: a system for synthesizing open-ended coding problems at scale. Starting from closed-ended coding tasks, FrontierSmith mutates, filters, and builds runnable optimization environments for long-horizon coding agents. In our experiments, FrontierSmith data trains stronger models than human-curated open-ended data on FrontierCS and ALE-bench. Blog: Paper: Code: Model:
0:56

Open-ended coding training data may no longer be the bottleneck: AI can scale open-ended tasks—and even outperform human-expert curation. FrontierCS team is releasing FrontierSmith: a system for synthesizing open-ended coding problems at scale. Starting from closed-ended coding tasks, FrontierSmith mutates, filters, and builds runnable optimization environments for long-horizon coding agents. In our experiments, FrontierSmith data trains stronger models than human-curated open-ended data on FrontierCS and ALE-bench. Blog: Paper: Code: Model:

Qiuyang Mang

122,932 görüntüleme • 6 ay önce

Announcing Apex: agentic coding model for mobile and web, specialized in React. We built Apex from a single idea: can an open-weight model deliver frontier-level performance at lower cost? Yes. Apex is SOTA in React Native and Next.js. Try it now:
1:47

Announcing Apex: agentic coding model for mobile and web, specialized in React. We built Apex from a single idea: can an open-weight model deliver frontier-level performance at lower cost? Yes. Apex is SOTA in React Native and Next.js. Try it now:

Callstack Engineers

52,813 görüntüleme • 1 ay önce