Hasta La Gadget!

Xiaomi has officially entered the artificial intelligence space with the release of its first open-source reasoning model, MiMo 7B. Developed by the company’s newly established Big Model Core Team, MiMo represents a strategic shift from Xiaomi’s traditional hardware focus to advanced AI research and development.

Despite having just 7 billion parameters, MiMo has demonstrated strong performance in complex reasoning tasks, surpassing much larger models like OpenAI’s o1-mini and Alibaba’s 32B QwQ-Preview. This performance has placed Xiaomi in the spotlight for producing a highly efficient and compact model capable of rivaling more resource-intensive competitors.

Benchmark Performance and Design Strategy

MiMo 7B has outperformed its peers on benchmarks such as AIME 24-25 (a test for mathematical reasoning) and LiveCodeBench v5 (a programming challenge dataset). This success is attributed to a well-structured development process consisting of both pre-training and post-training innovations.

Pre-training Strategies Included:

• Rich Reasoning Corpus: Focused on extracting and integrating complex reasoning knowledge.
• Synthetic Data Generation: Produced approximately 200 billion tokens of expert-level reasoning data to improve training depth.
• Progressive Difficulty Training: Employed a three-phase training method with increasing levels of difficulty.
• Extensive Token Exposure: Trained across a vast 25 trillion tokens, ensuring comprehensive learning.

Post-training Enhancements:

• Test Difficulty-Driven Rewards: Introduced a novel approach for addressing reward sparsity in complex algorithmic tasks.
• Data Re-sampling Techniques: Applied to stabilize reinforcement learning processes.
• Seamless Rollout System: Increased training efficiency by 2.29 times and validation speed by 1.96 times, streamlining the reinforcement learning (RL) pipeline.

A Broader Vision for AI

The release of MiMo underscores Xiaomi’s broader ambitions in the AI sector. While previously known for consumer electronics and smart devices, Xiaomi is now positioning itself as a serious player in AI model development. By open-sourcing MiMo, Xiaomi is contributing to the collaborative AI research ecosystem, fostering innovation beyond proprietary development.

Developers and AI researchers can now access MiMo 7B and its full technical documentation via Xiaomi’s official Hugging Face repository, offering a valuable tool for further experimentation and development. This move reflects Xiaomi’s intent to build a strong presence in AI while supporting the open-source community.

MiMo 7B: Xiaomi’s Open Source AI Is A Tough New Rival To ChatGPT

, original content from Ubergizmo. Read our Copyrights and terms of use.

Xiaomi has officially entered the artificial intelligence space with the release of its first open-source reasoning model, MiMo 7B. Developed by the company’s newly established Big Model Core Team, MiMo represents a strategic shift from Xiaomi’s traditional hardware focus to advanced AI research and development.

Despite having just 7 billion parameters, MiMo has demonstrated strong performance in complex reasoning tasks, surpassing much larger models like OpenAI’s o1-mini and Alibaba’s 32B QwQ-Preview. This performance has placed Xiaomi in the spotlight for producing a highly efficient and compact model capable of rivaling more resource-intensive competitors.

Benchmark Performance and Design Strategy

MiMo 7B has outperformed its peers on benchmarks such as AIME 24-25 (a test for mathematical reasoning) and LiveCodeBench v5 (a programming challenge dataset). This success is attributed to a well-structured development process consisting of both pre-training and post-training innovations.

Pre-training Strategies Included:

• Rich Reasoning Corpus: Focused on extracting and integrating complex reasoning knowledge.
• Synthetic Data Generation: Produced approximately 200 billion tokens of expert-level reasoning data to improve training depth.
• Progressive Difficulty Training: Employed a three-phase training method with increasing levels of difficulty.
• Extensive Token Exposure: Trained across a vast 25 trillion tokens, ensuring comprehensive learning.

Post-training Enhancements:

• Test Difficulty-Driven Rewards: Introduced a novel approach for addressing reward sparsity in complex algorithmic tasks.
• Data Re-sampling Techniques: Applied to stabilize reinforcement learning processes.
• Seamless Rollout System: Increased training efficiency by 2.29 times and validation speed by 1.96 times, streamlining the reinforcement learning (RL) pipeline.

A Broader Vision for AI

The release of MiMo underscores Xiaomi’s broader ambitions in the AI sector. While previously known for consumer electronics and smart devices, Xiaomi is now positioning itself as a serious player in AI model development. By open-sourcing MiMo, Xiaomi is contributing to the collaborative AI research ecosystem, fostering innovation beyond proprietary development.

Developers and AI researchers can now access MiMo 7B and its full technical documentation via Xiaomi’s official Hugging Face repository, offering a valuable tool for further experimentation and development. This move reflects Xiaomi’s intent to build a strong presence in AI while supporting the open-source community.

MiMo 7B: Xiaomi’s Open Source AI Is A Tough New Rival To ChatGPT

, original content from Ubergizmo. Read our Copyrights and terms of use.

Xiaomi has officially entered the artificial intelligence space with the release of its first open-source reasoning model, MiMo 7B. Developed by the company’s newly established Big Model Core Team, MiMo represents a strategic shift from Xiaomi’s traditional hardware focus to advanced AI research and development.

Despite having just 7 billion parameters, MiMo has demonstrated strong performance in complex reasoning tasks, surpassing much larger models like OpenAI’s o1-mini and Alibaba’s 32B QwQ-Preview. This performance has placed Xiaomi in the spotlight for producing a highly efficient and compact model capable of rivaling more resource-intensive competitors.

Benchmark Performance and Design Strategy

MiMo 7B has outperformed its peers on benchmarks such as AIME 24-25 (a test for mathematical reasoning) and LiveCodeBench v5 (a programming challenge dataset). This success is attributed to a well-structured development process consisting of both pre-training and post-training innovations.

Pre-training Strategies Included:

• Rich Reasoning Corpus: Focused on extracting and integrating complex reasoning knowledge.
• Synthetic Data Generation: Produced approximately 200 billion tokens of expert-level reasoning data to improve training depth.
• Progressive Difficulty Training: Employed a three-phase training method with increasing levels of difficulty.
• Extensive Token Exposure: Trained across a vast 25 trillion tokens, ensuring comprehensive learning.

Post-training Enhancements:

• Test Difficulty-Driven Rewards: Introduced a novel approach for addressing reward sparsity in complex algorithmic tasks.
• Data Re-sampling Techniques: Applied to stabilize reinforcement learning processes.
• Seamless Rollout System: Increased training efficiency by 2.29 times and validation speed by 1.96 times, streamlining the reinforcement learning (RL) pipeline.

A Broader Vision for AI

The release of MiMo underscores Xiaomi’s broader ambitions in the AI sector. While previously known for consumer electronics and smart devices, Xiaomi is now positioning itself as a serious player in AI model development. By open-sourcing MiMo, Xiaomi is contributing to the collaborative AI research ecosystem, fostering innovation beyond proprietary development.

Developers and AI researchers can now access MiMo 7B and its full technical documentation via Xiaomi’s official Hugging Face repository, offering a valuable tool for further experimentation and development. This move reflects Xiaomi’s intent to build a strong presence in AI while supporting the open-source community.

MiMo 7B: Xiaomi’s Open Source AI Is A Tough New Rival To ChatGPT

, original content from Ubergizmo. Read our Copyrights and terms of use.

Nintendo is going after the company that showed off unauthorized mockups of the not-yet-announced Switch 2 at CES 2025. In a lawsuit filed on Friday, Nintendo accuses accessory maker Genki of trademark infringement, unfair competition and false advertising, arguing that Genki “embarked upon a strategic campaign intended to capitalize on the public interest surrounding Nintendo’s next-generation console.” According to Nintendo, Genki promoted accessories it claimed are compatible with the Switch 2 despite not having official access to the console, and used the Switch logo to do so.

Following the initial reports in January that Genki brought Switch 2 mockups to CES and published a video said to show a 3D render of the console, Nintendo released a statement to say that neither the hardware nor imagery were official, per IGN. While Genki reportedly claimed it had access to a Switch 2 at first, the company later wrote on X that it does “not own or possess a black market console,” and told Game*Spark (translated by Automaton) that its mockups were based solely on leaked information. In the lawsuit, Nintendo says Genki’s “statements were contradictory and inconsistent,” yet the company “maintained its representation to consumers that its accessories will be compatible with the Nintendo Switch 2 upon the console’s release.”

Nintendo officially announced the Switch 2 on January 16, a little over a week after the mockups made their CES appearance. Even in the months since, Nintendo alleges, Genki “continues to exploit the attention of the media and fan base and usurp and threaten sales of Nintendo’s own accessories,” through means including piggybacking on Nintendo Direct with its own “Genki Direct” event immediately after. In a statement shared on X following the filing, Genki said it is “taking [the lawsuit] seriously and working with legal counsel to respond thoughtfully.”

“What we can say is this: Genki has always been an independent company focused on building innovative gaming accessories for the community we love,” Genki’s post continued. “We’re proud of the work we’ve done, and we stand by the quality and originality of our products. While we can’t comment in detail, we’re continuing preparations to fulfill orders and showcase our newest products at PAX East this week.”

The Nintendo Switch 2 arrives after years of rumors and speculation. In the lawsuit, Nintendo notes that it began planning the next-gen console shortly after the release of the first. Pre-orders for the Nintendo Switch 2 opened in April, and the console will be available on June 5.

This article originally appeared on Engadget at https://www.engadget.com/gaming/nintendo/nintendo-is-suing-the-accessory-brand-that-flaunted-switch-2-mockups-ahead-of-its-reveal-210341646.html?src=rss

Nintendo is going after the company that showed off unauthorized mockups of the not-yet-announced Switch 2 at CES 2025. In a lawsuit filed on Friday, Nintendo accuses accessory maker Genki of trademark infringement, unfair competition and false advertising, arguing that Genki “embarked upon a strategic campaign intended to capitalize on the public interest surrounding Nintendo’s next-generation console.” According to Nintendo, Genki promoted accessories it claimed are compatible with the Switch 2 despite not having official access to the console, and used the Switch logo to do so.

Following the initial reports in January that Genki brought Switch 2 mockups to CES and published a video said to show a 3D render of the console, Nintendo released a statement to say that neither the hardware nor imagery were official, per IGN. While Genki reportedly claimed it had access to a Switch 2 at first, the company later wrote on X that it does “not own or possess a black market console,” and told Game*Spark (translated by Automaton) that its mockups were based solely on leaked information. In the lawsuit, Nintendo says Genki’s “statements were contradictory and inconsistent,” yet the company “maintained its representation to consumers that its accessories will be compatible with the Nintendo Switch 2 upon the console’s release.”

Nintendo officially announced the Switch 2 on January 16, a little over a week after the mockups made their CES appearance. Even in the months since, Nintendo alleges, Genki “continues to exploit the attention of the media and fan base and usurp and threaten sales of Nintendo’s own accessories,” through means including piggybacking on Nintendo Direct with its own “Genki Direct” event immediately after. In a statement shared on X following the filing, Genki said it is “taking [the lawsuit] seriously and working with legal counsel to respond thoughtfully.”

“What we can say is this: Genki has always been an independent company focused on building innovative gaming accessories for the community we love,” Genki’s post continued. “We’re proud of the work we’ve done, and we stand by the quality and originality of our products. While we can’t comment in detail, we’re continuing preparations to fulfill orders and showcase our newest products at PAX East this week.”

The Nintendo Switch 2 arrives after years of rumors and speculation. In the lawsuit, Nintendo notes that it began planning the next-gen console shortly after the release of the first. Pre-orders for the Nintendo Switch 2 opened in April, and the console will be available on June 5.

This article originally appeared on Engadget at https://www.engadget.com/gaming/nintendo/nintendo-is-suing-the-accessory-brand-that-flaunted-switch-2-mockups-ahead-of-its-reveal-210341646.html?src=rss