Photo by Steve Johnson on Unsplash
New AI Company Deep Cogito Releases First Hybrid AI Models
- Written by Andrea Miliani Former Tech News Expert
- Fact-Checked by Sarah Frazier Former Content Manager
The new American AI company Deep Cogito released its first open large language models (LLMs) called Cogito v1 on Tuesday. The startup claims its open-source and hybrid models outperform similar open AI models, including DeepSeek, Qwen, and Meta’s LLaMA.
In a rush? Here are the quick facts:
- Deep Cogito released its first LLM called Cogito v1 this Tuesday.
- The new American company claims its model outperforms equivalent open source models, including Qwen, DeepSeek, and Meta’s LLaMA.
- Cogito v1 models function in reasoning and standard mode, and have been optimized for function calling, coding, and agentic use.
Deep Cogito, headquartered in San Francisco, introduced Cogito v1 through an official announcement on its website, including details of its LLMs, its performance, and upcoming releases.
“We are releasing the strongest LLMs of sizes 3B, 8B, 14B, 32B, and 70B under open license,” states the document. “Each model outperforms the best available open models of the same size, including counterparts from LLaMA, DeepSeek, and Qwen, across most standard benchmarks. In particular, the 70B model also outperforms the newly released Llama 4 109B MoE model.”
The new model has been trained with the Iterated Distillation and Amplification (IDA), a framework that trains AI models to build aligned systems for general superintelligence, and includes “reasoning” features. All models can function in reasoning and standard mode—a hybrid modality—and have been optimized for function calling, coding, and agentic use.
According to the startup’s research and results, Cogito v1 outperformed most equivalent models in direct and reasoning modes, considering popular benchmarks. The company expects to launch larger models in the next few months.
“Our next release will feature updated checkpoints for each model size (3B to 70B), with extended training periods, as well as larger models in the coming weeks and months,” wrote the company. “All models will be open source.”
Cogito v1 can be downloaded on Ollama or Hugging Face, or accessed through APIs on Together AI or Fireworks AI.
Cogito v1 Preview is now live on Together AI ✨ These open models from Deep Cogito push the boundaries of reasoning and alignment, with options up to 70B parameters. Available now via Together Dedicated Endpoints. pic.twitter.com/hc0gYXU7x1 — Together AI (@togethercompute) April 8, 2025
According to TechCrunch , Deep Cogito was founded in June 2024 by Drishan Arora and Dhruv Malhotra—both former Google employees. The company is backed by South Park Commons, and its main goal is to build a general superintelligence that can outperform humans.
The new AI model arrives to keep up with the intensity and speed of the AI market and to join the open-source trend. The Chinese search engine giant Baidu announced a new upcoming open-source AI model in February, and DeepSeek shared more details and a transparency initiative for its code just a few weeks ago.
Image by DC Studio, from Freepik
Scattered Spider Evolves in 2025 with New Phishing Kit and Malware
- Written by Kiara Fabbri Former Tech News Writer
- Fact-Checked by Sarah Frazier Former Content Manager
The notorious hacking group Scattered Spider continues to pose a serious cybersecurity threat in 2025, despite multiple arrests in the past year.
In a rush? Here are the quick facts:
- Spectre RAT malware updated for stealthy, long-term system access.
- Group targets brands like Nike, T-Mobile, and Pure Storage.
- Rentable subdomains and recycled domains complicate threat tracking.
The group uses sophisticated social engineering tactics but has evolved its methods by introducing new phishing kits, and an updated Spectre RAT malware to attack high-profile companies.
According to cybersecurity firm Silent Push , Scattered Spider remains actively engaged in attacks on major brands including Nike, T-Mobile, Louis Vuitton, and Vodafone. They’ve also expanded their targets to include cloud storage and marketing platforms such as Pure Storage and Klaviyo.
Since 2022, the group has been active and initially became known for breaking into companies such as Twilio and MGM Resorts. It did so by deceiving employees into giving away their login credentials and MFA codes via fake login portals.
Although several members, including the alleged leader Tyler Buchanan were arrested in 2024, the group has since come back to life, likely with new members and developers improving their tools and techniques, as explained by Silent Push.
One of the most notable evolutions this year is their adoption of Phishing Kit #5, now hosted on Cloudflare. Silent Push explains that the current version differs from earlier versions which redirected users to Rick Astley’s “Never Gonna Give You Up” as a joke because it operates more discreetly and is harder to detect.
In another troubling shift, the group has started leveraging publicly rentable subdomains—such as klv1.it[.]com—that mimic legitimate services. These subdomains, often tied to dynamic DNS providers, are harder to trace due to their lack of traditional domain registration.
Silent Push warns that organizations should consider blocking such domains at the network level to reduce exposure.
Additionally, Scattered Spider has been linked to the reacquisition of a domain once owned by Twitter/X: twitter-okta[.]com. While it remains uncertain whether the domain will be used in upcoming campaigns, it underscores the group’s persistence in exploiting overlooked or abandoned digital assets, says Silent Push.
The Scattered Spider group continues to evolve as a dangerous threat in 2024 because of their ability to adapt their infrastructure and malware while finding new attack vectors. The group’s ongoing evolution shows they have not completed their operations.
Organizations need to stay vigilant while tracking unusual behavior and maintain updated security measures to prevent attacks from this persistent cybercriminal organization.
