MiniMax has announced MiniMax M3, promoting it as a breakthrough "open-weights" model that integrates a 1M token context window, native multimodality, and advanced coding and agentic capabilities, despite current community skepticism over the actual availability of the model's weights.

// 1h agoMODEL RELEASE

MiniMax has announced MiniMax M3, promoting it as a breakthrough "open-weights" model that integrates a 1M token context window, native multimodality, and advanced coding and agentic capabilities, despite current community skepticism over the actual availability of the model's weights.

MiniMax has announced MiniMax M3, which is marketed as the first 'open-weights' model to combine three frontier capabilities: a million-token context window powered by the proprietary MiniMax Sparse Attention (MSA) architecture, native multimodal reasoning trained from the ground up, and state-of-the-art coding and agentic capabilities designed for executing complex, long-horizon tasks. While the model represents a major advancement in context efficiency and autonomous performance—exhibiting the ability to reproduce research papers without human intervention—the community has expressed skepticism due to the lack of publicly accessible model weights. Currently, developers can only access the model's capabilities through MiniMax's API or platform partners like Ollama, rather than downloading the weights for local deployment.

// ANALYSIS

Calling MiniMax M3 'open-weights' while keeping the actual weights locked behind an API is a deceptive marketing stunt that highlights the growing industry trend of 'open-washing' proprietary models.

* **Proprietary Sparse Attention:** The underlying MiniMax Sparse Attention (MSA) architecture is highly efficient, allowing sub-quadratic processing of up to 1 million tokens, representing a strong technical leap.

* **Native Multimodality:** Training natively on multimodal data from the start ensures stronger cross-modal semantic alignment than standard vision-adapted models.

* **Agentic Prowess:** The model is optimized for complex software engineering and long-horizon tasks, including the autonomous reproduction of research papers.

* **Open-Washing Backlash:** Despite the open-weight branding, the lack of downloadable weights, parameter counts, or training code has drawn significant skepticism from the open-source developer community.

// TAGS

minimaxminimax-m3llmopen-weightsagentmultimodal1m-contextai

DISCOVERED

1h ago

2026-06-01

PUBLISHED

1h ago

2026-06-01

RELEVANCE

8/ 10

AUTHOR

DeepInfra

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

OPEN SOURCE54m ago

parakeet.cpp delivers 2x faster local ASR

Developed by the LocalAI team, parakeet.cpp is a dependency-free C++17 inference engine for NVIDIA's NeMo Parakeet ASR models that runs up to 2x faster than standard baselines. By leveraging the ggml library to eliminate Python runtime dependencies, it enables highly portable offline speech recognition across CPUs and multiple GPU backends.

MODEL1h ago

NVIDIA has launched Nemotron 3 Ultra, a massive 550-billion parameter hybrid Mamba-Transformer model that is five times faster and 30% cheaper than comparable open models.

NVIDIA has unveiled Nemotron 3 Ultra, the flagship model in its new Nemotron 3 family of foundation models built specifically for complex reasoning, planning, and multi-step agentic workflows. Featuring 550 billion total parameters, the model is powered by a hybrid Mamba-Transformer Mixture-of-Experts (MoE) architecture that utilizes LatentMoE token compression and NVFP4 training to activate only 55 billion parameters per token. This efficient design enables throughputs exceeding 300 tokens per second, a 1-million token context window, and up to 5x faster inference at 30% lower cost compared to existing open-weights models in its class.

VIDEO1h ago

Socket Firewall blocks dependency malware in real time by intercepting package manager network requests at install time.

Socket Firewall is a free command-line interface (CLI) tool designed to proactively protect developer environments from malicious software supply chain attacks. By acting as a local network proxy or command prefix, the tool intercepts package manager network requests (supporting npm, pip, and cargo) during installation, evaluating dependencies in real-time against Socket's security API. This allows it to automatically block known malware, flag suspicious packages with risky capabilities (like unexpected telemetry, network, or filesystem access), and enforce safety policies without disrupting developer workflows.