MAI Models vs Step 3.7 Flash

Microsoft CEO emphasizes shifting from single large models to ecosystem-driven full-stack construction, enabling enterprises to participate AI-native; proposes building AI stacks (tools, data, governance) and governance frameworks; announces MAI models with strong pre-training and explainability, underpinned by Azure’s 15-month capacity growth exceeding past 15 years.

Microsoft positions itself as the Frontier Intelligence Platform, with enterprises creating value beyond Microsoft by leveraging OpenClaw/Scout, Work IQ, and private evaluations/traces as Token IP. The platform is defined by its ability to create more value than what is captured in the platform, with a training focus on clean lineage, cognitive cores, and MAI models that let 5B models climb performance in real tasks.

StepFun released Step 3.7 Flash — a high-efficiency agentic coding model supporting multimodal understanding, tool use, and long-running workflows; its standout feature is full free access in Hermes Agent, removing typical API/credit barriers for real-world testing.

Step 3.7 Flash by Yujue Star is a new-generation Flash model for production-grade AI Agents, featuring native multimodal understanding, high throughput with low latency, and enhanced web search. It achieves 97% of Claude Opus 4.6's coding performance at only 1/9 the cost per task, ideal for high-frequency, complex real-world workflows.

Step 3.7 Flash is a 196B MoE model designed from the ground up for inference efficiency, using MFA and AFD techniques to reduce KV-cache usage to ~22% of DeepSeek, supporting agent, coding, and multimodal workflows, open-sourced under Apache 2.0 and available on Fireworks.

NVIDIA introduces Step 3.7 Flash for large-scale MoE models.

文章仅提供了一个链接，介绍如何在 NVIDIA GPU 上运行 Step 3.7 Flash，缺乏深度和技术细节。

Step 3.7 Flash significantly reduces KV-cache cost via MFA + AFD technology, enabling efficient inference with one-click deployment.

Step 3.7 Flash model achieves 400 tokens/sec processing speed, reducing Agent task cost to a fraction of Claude's.