MiniMax (official) (@MiniMax

MiniMax (official)

1,617 posts

MiniMax (official)

@MiniMax_AI

Agent: @MiniMaxAgent Token Plan: platform.minimax.io/subscribe/toke… API: platform.minimax.io Community: minimax.io/discord

San Francisco

Joined January 2025

Pinned
MiniMax (official)
@MiniMax_AI
Jun 1
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 -
5.2M
MiniMax (official)
@MiniMax_AI
9h
More great options for the open-weight ecosystem. Thanks @NVIDIAAI for making MiniMax M3 available in NVFP4.
NVIDIA AI
@NVIDIAAI
10h
Replying to @QuantCapitalX and @MiniMax_AI
huggingface.co/nvidia/MiniMax…
17K
MiniMax (official)
@MiniMax_AI
17h
we’ll be at AI Engineer After Dark on July 1st with @vercel , @merge_api, @FactoryAI , and a room full of people building the AI engineering stack. our Research Lead, RL Training @olive_jy_song will be giving a lightning talk on post-training MiniMax M3 as part of a lineup
8.3K
MiniMax (official)
@MiniMax_AI
16h
Register here ->
AI Engineer After Dark · Luma
From luma.com
2.1K
MiniMax (official) reposted
Creao AI
@CreaoAI
Jun 24
Last night in SF, we gave a room full of creators one challenge: come with an idea, leave with a working demo. Our CTO @intuitiveml opened with a talk on what it actually takes to build AI-first. Agents that run on loops and improve with every run. A harness that makes the whole
5.6K
MiniMax (official)
@MiniMax_AI
Jun 24
M3 is now the default builder model in Kimchi Coding by @cast_ai not every coding task should go through the same model. Kimchi routes work based on complexity, cost, and deployment needs. that makes it a natural fit for M3: open weights, 1M context, and strong coding
11K
MiniMax (official)
@MiniMax_AI
Jun 24
Read the story here ->
Cast AI adds MiniMax M3 to Kimchi Coding as default model
From itbrief.co.uk
2.7K
MiniMax (official)
@MiniMax_AI
Jun 24
research -> synthesize -> build. fun example of using M3 and @opencode to learn about a new domain instead of reading 20 tabs
Amber Shen
@whosamberella
Jun 24
I wanted to learn about humanoid robots 🤖, then decided the most reasonable thing to do was build myself a humanoid catalog to learn from. humanoid-index-one.vercel.app @MiniMax_AI M3 researched the space, then built the catalog in @opencode . The SVG prototype doesn’t just look
00:00
10K
MiniMax (official)
@MiniMax_AI
Jun 23
Exciting conversations happening on July 1st with @friendliai @NVIDIAAI @kilocode
friendliai
@friendliai
Jun 23
Lineup for Inside The AI Coding Stack (7/1): @nvidia — Harry Kim on GPU infra for AI-native workloads @friendliai — Gon Chun on frontier AI inference for agents @MiniMax_AI — Victor Su-Ortiz on M3 + reasoning @kilocode — Brian Turcotte on agentic coding in production Spots are
7.5K
MiniMax (official)
@MiniMax_AI
Jun 23
Replying to @MiniMax_AI @hud_evals and @ycombinator
Tera — zero-token browser use solve once, reuse forever: a browser RL env that learns from past traces, then replays matched workflows with 0 LLM tokens. took #3 overall.
4K
MiniMax (official)
@MiniMax_AI
Jun 23
Replying to @MiniMax_AI @hud_evals and @ycombinator
Warehouse AI an RL environment for autonomous warehouse robot fleets: fulfill orders, avoid collisions, coordinate dispatch, and learn before touching the real world.
2.1K
MiniMax (official)
@MiniMax_AI
Jun 23
some scenes from the @hud_evals × @ycombinator Frontier RL Environments Hackathon over the weekend. 24 intense, collaborative hours with 200 of the top builders in the world turning browser-use, robotics, logistics, and worldsims into environments where models can actually learn
7.1K
MiniMax (official)
@MiniMax_AI
Jun 23
Atomz autonomous Mars construction rovers in simulation — tracked robots navigating terrain, scanning sites, and 3D-printing habitat structures from regolith.
2.1K
MiniMax (official)
@MiniMax_AI
Jun 23
Always fun to see frontier models tested in hackathon conditions. Looking forward to the projects with @SinguarityNET and free M3 through @S_Compute
SingularityNET
@SingularityNET
Jun 22
We are pleased to announce that @MiniMax_AI, a global foundation model company, is sponsoring the BGI Sprint hackathon, supporting participants in building with the ASI Alliance technology stack and MiniMax's models. MiniMax's proprietary models and AI-native products serve over
00:00
13K
MiniMax (official) reposted
Cerebral Valley
@cerebral_valley
Jun 22
The second official Engineer World’s Fair Hackathon with @aiDotEngineer Join us for 48 hours of building at the frontier of recursive self-improvement (RSI): systems that learn from their own outputs and become more capable through continuous iteration. Excited to be partnering
15K
MiniMax (official)
@MiniMax_AI
Jun 22
M3's 95% cache ratio tells the more interesting story hint: long-horizon coding loops appreciate @opencode making this usage visible.
Jay
@jayair
Jun 21
The other open source models are clawing market share back from DeepSeek
32K