Xiaomi has launched its new open-source artificial intelligence (AI) model, MiMo-V2-Flash, aimed at competing with established offerings from tech giants such as DeepSeek, Moonshot AI, Anthropic, and OpenAI. The model is accessible globally via Xiaomi’s developer platform, MiMo Studio, in addition to platforms like Hugging Face and API Platform. According to Xiaomi, MiMo-V2-Flash excels in reasoning, coding, and agentic scenarios, presenting itself as a versatile assistant for everyday tasks.
The announcement gains additional significance as Luo Fuli, a notable AI expert who recently transitioned from DeepSeek to Xiaomi’s MiMo team, described MiMo-V2-Flash as “step 2 on our AGI road map.” AGI, or artificial general intelligence, is a conceptual form of AI that aspires to match or surpass human cognitive abilities. This remark underscores Xiaomi’s ambitions within the advanced AI development landscape.
Xiaomi has touted MiMo-V2-Flash as a cost-effective and high-performance AI model, claiming it was designed for maximum efficiency. The company states that it delivers “blazing-fast inference at 150 tokens per second” while maintaining an ultra-low cost of $0.1 per million input tokens and $0.3 per million output tokens. This pricing structure positions it as one of the most affordable high-performance models currently available, according to Xiaomi.
In terms of performance, Xiaomi asserts that MiMo-V2-Flash matches or exceeds the capabilities of Moonshot AI’s Kimi K2 Thinking and DeepSeek’s V3.2 Thinking across most reasoning benchmarks. The company further claims that its new model outperforms Kimi K2 Thinking in evaluations involving long-context inputs. On agentic tasks, MiMo-V2-Flash achieved a score of 73.4% on SWE-Bench Verified, surpassing all open-source competitors and coming close to the performance levels of OpenAI’s GPT-5-High.
The introduction of MiMo-V2-Flash signifies Xiaomi’s strategic intent to carve out a significant space in the rapidly evolving AI market. The competition is intensifying as companies race to develop models that not only perform well but also do so at a lower cost. Xiaomi’s pricing and performance claims could attract developers and businesses looking for effective AI solutions in a cost-sensitive environment.
As AI technology continues to advance, the implications of Xiaomi’s foray into open-source AI could be far-reaching, potentially democratizing access to high-performance AI tools. The market’s response to MiMo-V2-Flash will likely influence how other companies position their offerings and strategies moving forward. Xiaomi’s development path towards AGI, as articulated by Fuli, suggests a long-term vision that extends beyond immediate market competition, aiming to reshape the future of AI capabilities.
See also
AI Model CXR-Age Outperforms DNA-Based Clocks in Detecting Heart, Lung Aging Risks
Google and Meta Collaborate on TorchTPU to Compete with Nvidia’s AI Chip Dominance
AI-Driven Workforce Changes Demand New Leadership and Skills, Says Sedgwick Report
Wall Street Sinks as AI Funding Fears Weigh on Tech Stocks; Oracle Drops 5.4%
Elon Musk’s xAI Launches Grok in El Salvador to Serve 1 Million Students by 2027
