Search results for Gemma
AI & Enterprise
Kakao language model Kanana-2 outperforms overseas models in safety assessment
Kakao has disclosed numerical safety verification results for its self-developed lightweight language models. The company said on Aug. 18 that two open-source models in its Kanana-2 series scored higher than major global models of similar size on assessments of harmfulness and bias, confirming top-tier safety. The evaluation used AssurAI, a Korea-designed safety benchmark, and compared the models with Google’s Gemma and Alibaba’s Qwen.
AI & Enterprise
China releases top AI for free in \'open-weight strategy\' targeting U.S. big tech
Chinese artificial intelligence firms are releasing high-performance open-weight models for free, putting pressure on U.S. big tech companies that rely on closed AI platforms. The Verge reported that Moonshot AI’s Kimi K3 is seen as matching some top U.S. models at lower cost and is widening the fight to control the developer ecosystem. Analysts say open weights can still generate revenue through infrastructure and services. The trend is also tied to Beijing’s industrial strategy and policy debates in the United States.
AI & Enterprise
Kakao releases lightweight AI model Kanana-2 as open source
Kakao has released four lightweight language models in its Kanana-2 series as open source on global AI platform Hugging Face, targeting on-device use such as smartphones. The company said the models deliver global-level performance in both Korean and English and outperform similar-size open-source models on major benchmarks. Kakao also applied its Korean-specialised tokenizer to boost Korean processing efficiency by more than 30 percent and introduced sliding window attention to cut memory use by up to 72.7 percent in long conversations.
-
AI & Enterprise
DeepSeek unveils open-source DSpark technology to boost AI inference speed by up to 85 percent
-
Industry
Hanwha Power enters U.S. fuel gas compressor market for first time
-
AI & Enterprise
KT unveils Korean culture-focused AI safety benchmark KSAFE-MM
-
AI & Enterprise
Google unveils Gemma 4 QAT for mobiles and laptops to cut AI memory use sharply
-
AI & Enterprise
Nvidia unveils 550 billion-parameter Nemotron 3 Ultra, starts mass production of Vera Rubin
-
AI & Enterprise
GitHub tool bypasses AI safeguards in some Meta, Google open-weight models, test finds
-
AI & Enterprise
Selectstar red-teaming technology adopted by ACL 2026
-
AI & Enterprise
Nebius to buy AI inference optimisation firm Eigen AI for $643 million
-
AI & Enterprise
Personalisation that general-purpose AI cannot do can be solved with local AI, founder says
-
AI & Enterprise
Why Google Gemma 4 is in the spotlight as an open-source AI game changer
-
AI & Enterprise
LG unveils EXAONE 4.5 multimodal AI model focused on stronger reasoning
-
AI & Enterprise
Using Google\'s offline dictation app Eloquent, it polished speech in real time
-
AI & Enterprise
Google quietly launches offline AI dictation app for iOS
-
AI & Enterprise
Evolving software pricing plans; Chinese AI firms revise strategy
-
AI & Enterprise
Google unveils open model Gemma 4, supports complex reasoning on low-power devices