MiniMax (official)

1.7K posts

MiniMax (official) banner
MiniMax (official)

MiniMax (official)

@MiniMax_AI

Agent: @MiniMaxAgent Token Plan: https://t.co/BDCycxepZw API: https://t.co/fHRdSV7BwZ Community: https://t.co/uhxxfLgkLU

San Francisco 加入时间 Ocak 2025
878 关注103.6K 粉丝
置顶推文
MiniMax (official)
MiniMax (official)@MiniMax_AI·
Introducing MiniMax M3: The First Open-Weights Model to Combine Three Frontier Capabilities - Coding & Agentic Frontier: 59.0% SWE-Bench Pro, 66.0% Terminal Bench 2.1, 34.8% SWE-fficiency, 28.8% KernelBench Hard, 74.2% MCP Atlas - MiniMax Sparse Attention scales context to 1M - Natively Multimodal from Step Zero API: platform.minimax.io Token Plan: platform.minimax.io/subscribe/toke… 🚀New! MiniMax Code: code.minimax.io Weights & Tech Report in ~10 Days
MiniMax (official) tweet media
English
569
1.1K
11.5K
5.2M
MiniMax (official)
MiniMax (official)@MiniMax_AI·
Paris made one thing clear: open models are no longer a side conversation, they’re becoming central to how frontier AI is built, and scaled. 🇫🇷 Across RAISE week: — Linda Sheng, President of Global Business at MiniMax, joined Luca Schmid, Managing Director HSG, on the Ada Lovelace Stage to discuss open models, and native multimodality, and the new enterprise use cases they make possible. — RAISE House brought @togethercompute, @cast_ai, and @magnific together for a panel on open frontier models, applications, and infrastructure. — Our private MiniMax Salon created space for deeper conversations with founders, developers, investors, and partners from @runware, @youdotcom, @ExaAILabs, and @FeatherlessAI — Our reception with @SambaNovaAI, we brought together builders and infrastructure partners—including @InfercomAI, Argyll, and SCX ai—above Paris for live demos and discussions. — At the booth, we met builders exploring how open models can move into production. Thank you to everyone who joined us. Until next time, Paris. Intelligence with Everyone. 🌍
MiniMax (official) tweet mediaMiniMax (official) tweet mediaMiniMax (official) tweet mediaMiniMax (official) tweet media
English
4
5
109
9.7K
MiniMax (official)
MiniMax (official)@MiniMax_AI·
Sparse attention only matters if the systems can preserve the theoretical gains. @FireworksAI_HQ new M3 kernel on Blackwell uses a KV-stationary design to read each selected block once, reaching ~980 TFLOP/s on a B200. Read the full breakdown below for a closer look at the kernel design and optimizations. 👇
Fireworks AI@FireworksAI_HQ

Long-context sparse attention has a catch: data-dependent block selection wrecks memory access kills speed. Our @MiniMax_AI M3 kernel on Blackwell answers it. KV-stationary, each block read once, ~980 TFLOP/s on a B200. See the breakdown here → fireworks.ai/blog/kernel-op…

English
9
15
282
27.6K
MiniMax (official) 已转推
MiniMax (official) 已转推
MiniMax (official) 已转推
vLLM
vLLM@vllm_project·
@RyanLeeMiniMax Congratulations to the @MiniMax_AI team. Excited to see this strong commitment to open source and the broader AI ecosystem. 🚀 Looking forward to building the future of AI together. 🤝
English
1
2
68
10.4K
MiniMax (official)
MiniMax (official)@MiniMax_AI·
Congrats to the @ollama team on such a massive milestone 🥳. The open model ecosystem continues to grow, and we’re proud to be building alongside the Ollama team as we make frontier AI more accessible. Here’s to expanding access to open models globally 🚀.
ollama@ollama

Big day for Ollama! When we started, open models and the open source AI ecosystem were in their early days with few believers. Our belief in open source has never wavered. With today's fundraising announcement and our 9M+ active builders, we’re ready to scale open models into AI that you can own. All aboard open models! 🧵

English
5
7
144
17.1K
ollama
ollama@ollama·
Big day for Ollama! When we started, open models and the open source AI ecosystem were in their early days with few believers. Our belief in open source has never wavered. With today's fundraising announcement and our 9M+ active builders, we’re ready to scale open models into AI that you can own. All aboard open models! 🧵
ollama tweet media
English
205
237
2K
292.2K
MiniMax (official)
MiniMax (official)@MiniMax_AI·
AdaL Engineer doesn’t just use one model for every part of the loop: the repeated cycle of planning, building, testing, and improving. Rather, it routes each stage to the model best suited for the task. For browser automation, the team turns to MiniMax M3, where strong visual understanding and computer-use capabilities are a natural fit. As agent systems become more complex, choosing the right model for each step may is becoming just as important as the agent harness.
Li Yin@panda_liyin

x.com/i/article/2074…

English
7
7
99
12.1K
Prime Intellect
Prime Intellect@PrimeIntellect·
Announcing our $130M Series A to build the Open Superintelligence Stack Led by Radical Ventures, with NVIDIA, Intel Capital, Dell Capital, and existing investors Train, deploy, and continuously improve your own models using our stack. Own your intelligence.
English
332
539
4.8K
1.3M
MiniMax (official)
MiniMax (official)@MiniMax_AI·
Excited to see this go live with the @togethercompute team. As more production workloads move to open models, reliable infrastructure becomes just as important as the models themselves. Proud to see M3 launch with Provisioned Throughput. 🤝
Together AI@togethercompute

We're introducing Provisioned Throughput: reserved inference capacity for frontier open models, with token-based pricing and a 99% uptime SLA. Serverless simplicity, guaranteed capacity, up to 90% lower cost vs. Opus 4.8. Get started with MiniMax M3 + GLM-5.2, read more 🧵

English
8
13
142
19.3K
MiniMax (official) 已转推
Kilo
Kilo@kilocode·
$200 to $20. That's how your AI costs will change if you switch from Claude Code Max to the @MiniMax_AI Token Plan Plus in Kilo. MiniMax's latest model, M3, performs at 79% of Sonnet 5's level for a tenth of the price, and the plan gets you 1.7B tokens per month through Kilo. See how much you'd save with model freedom: kilo.ai/calculator
English
11
20
181
25.3K
Hermes Agent Tips
Hermes Agent Tips@HermesAgentTips·
if you had to pick out of these 3 models for hermes agent which one you think is best and why - DeepSeek V4 Flash - MiMo-V2.5 - MiniMax M3
English
124
2
141
32.6K
MiniMax (official)
MiniMax (official)@MiniMax_AI·
July 9 - RAISE House with MiniMax. A private executive gathering with: Rio Shen, GM EMEA at MiniMax Kai Mak, CRO at @togethercompute Laurent Gil, Co-Founder & President at @cast_ai Claire Xue, Global Head of AI Ecosystem at @magnific Expect a conversation on frontier multimodal AI in real-world deployment. luma.com/h1emgp78
MiniMax (official) tweet media
English
0
0
45
4.2K
MiniMax (official)
MiniMax (official)@MiniMax_AI·
MiniMax is heading to RAISE Week in Paris 🇫🇷 Find us at Booth 32D in Delorme, in a fireside chat on the Ada Lovelace Stage, and at our private executive gathering: RAISE House with MiniMax. We’ll be talking M3, open weights, multimodal AI, and where frontier AI is moving next — including how MiniMax is helping shape it. Details below 👇
English
9
10
157
18.6K