Gemini 4 Argon é stato appena annunciato da Google e rappresenta a detta loro il nuovo modello AI
Google announces Gemini 4 Argon, a new AI model claimed to compete with OpenAI and Anthropic offerings, featuring strong performance across benchmarks and enhanced security against prompt injection attacks. The model will be released through a staged rollout starting with the Fairwind Program for selected companies, followed by API access for developers, and eventually public release.
Summary
Google has announced Gemini 4 Argon, positioned as a competitive response to OpenAI and Anthropic's latest models. According to published benchmarks, the model outperforms GPT-6 Astra, Fable 5.1, and Opus 5, establishing it as state-of-the-art. The model demonstrates particular strength in programming, physics, chemistry, and mathematics. A key differentiator is its enhanced security, specifically its resistance to prompt injection attacks, contrasted with Chinese models which are described as significantly weaker and more vulnerable to such attacks. The release strategy follows a measured approach: initially, only selected companies will gain access through Google's Fairwind Program. Subsequently, the model will be made available via API for developers, and finally released to the general public once quality and security testing is completed. Pricing is set at $2 per million input tokens and $10 per million output tokens. On independent AI benchmarks, Gemini 4 Argon positions itself between GPT-6 Astra and GPT-6.1 Sol. The speaker acknowledges the increasing difficulty in evaluating competing models, noting that different use cases will favor different models. The announcement represents Google's return to competitive positioning after a prolonged period, allowing the company to compete alongside Chinese AI developers, Anthropic, and OpenAI. The speaker personally notes the relevance of having €300,000 in Google Cloud credits available for use with this new model, credits provided for his London-based startup Atom with support from investment funds.
Key Insights
- Gemini 4 Argon is claimed to beat GPT-6 Astra, Fable 5.1, and Opus 5 according to published benchmarks, establishing state-of-the-art performance
- The model's key security advantage is resistance to prompt injection attacks, with Chinese models specifically characterized as much weaker and easier to attack by comparison
- Google is implementing a three-stage release strategy: first through the Fairwind Program for selected companies, then via API for developers, and finally to the public after quality and security testing
- On independent AI benchmarks, Gemini 4 Argon positions itself halfway between GPT-6 Astra and GPT-6.1 Sol, with the speaker noting increasing difficulty in evaluating competing models across different use cases
- The speaker frames Google's announcement as the company finally returning to competitive positioning on the podium after a long wait, now able to compete with Chinese developers, Anthropic, and OpenAI
Topics
Transcript
[0:00] Google finally releases a noteworthy new AI model that can compete against Open AI and Anthropic. It's called Gemini 4 Argon. From published benchmarks it seems to beat GPT6 Astra, Fable 5.1, Opus 5. State of the art. It is an excellent agent, very strong in programming, physics, chemistry, mathematics, but above all it is much more secure because it resists prompt injection attacks, unlike the [0:30] Chinese models which are much weaker and easier to attack. The model has n't been released yet, it's just been announced and they've created their own program called the Fairwind Program, where only selected companies will be able to access it to start using it, then they'll release this model under the…
Full transcript available for MurmurCast members
Sign Up to AccessMore from Simone Rizzo
OpenAI dots compete con grokbot il nuovo gpt 6.1 sol compete con opus 5.5 tutto questo è successo
OpenAI released multiple new frontier models including GPT 6.1 Sol at significantly lower token costs, along with cloud-based agents called Dots that compete with Grockbot. The company introduced new subscription tiers, collaborative spaces, a Decisions API for real-time decision-making, and cloud-based coding tools, while Elon Musk strategically acquired dots.com to redirect traffic to Grockbot.
OpenAI copia Grok con i dots? GPT-6.1 sfida Opus 5.5
A comprehensive analysis of OpenAI's DevDay announcements, including new AI agents called 'Dots,' the GPT-6.1 model, and pricing changes, alongside broader trends in AI development from competitors like Anthropic and open-source projects. The speaker critiques the announcements as largely incremental improvements rather than groundbreaking innovations.
Apple rilascia un nuovo modello AI si chiama LensVLM e cerca di risolvere il problema dei token in
Apple releases LensVLM, a fine-tuned 9 billion parameter model based on Qwen 3.5, implementing a novel technique called Selective Context Expansion that converts text into images to achieve 15x token compression. This approach allows visual language models to handle large text inputs more efficiently while maintaining reasoning accuracy.
Usare Jev per risparmiare tantissimi token nel tuo agente AI mantenendo alte le prestazioni.
Tamara discovered a method to use JEV, a fast decision-making model, to replace the slow slash compact function in AI agents. By having JEV decide whether to keep or discard function outputs, the system clears context window space much faster while maintaining performance. The solution, called Fast Jeev Comp, has gained significant popularity and can also be run locally using open-source models.
Ti svelo il miglior modello AI locale da far girare sul tuo pc con 4Miliardi di parametri.
Spark X 2.5 is highlighted as the best 4-billion parameter local AI model offering an optimal balance between speed and performance. The model outperforms competitors in its size class, matches the performance of 9B models, and can handle up to 1 million input tokens, making it suitable for various local AI applications.