OpenAI releases GPT-4.1!

OpenAI releases GPT-4.1!

OpenAI today released three new models that could revolutionize the AI landscape: GPT-4.1, GPT-4.1 mini and GPT-4.1 nano. These new models outperform their predecessors in all areas and offer dramatic improvements in programming and following instructions.

The advances in coding are particularly impressive: GPT-4.1 achieves 54.6% on the SWE-bench Verified Benchmark - an increase of 21.4% over GPT-4o. For long contexts, the models can now process and understand up to 1 million tokens. That's like eight complete React codebases in one go!

The new models are not only more powerful, but also more cost-effective. GPT-4.1 mini even outperforms GPT-4o in many benchmarks, while cutting latency in half and reducing costs by 83%.

Particularly exciting for developers: the improved reliability in instruction following and understanding of long contexts makes these models ideal for “agents” - systems that can independently perform complex tasks on behalf of users.

In addition to GPT-4.1, the release of full o3, the open source model and o4-mini is expected this week. This week is hot!

The global AI robot market will generate billions in revenue.

Wayve and Nissan are cooperating on autonomous driving

British AI startup Wayve, backed by investors including Nvidia, SoftBank and Microsoft, is partnering with Nissan. Starting in 2027, Wayve's autonomous driving software will be integrated into Nissan's ProPILOT system. Unlike other companies, Wayve relies on camera-based AI that adapts like a human driver, which has the potential to significantly advance the development of autonomous vehicles.

GXO tests humanoid robots in logistics

Logistics giant GXO is piloting the use of humanoid robots from companies such as Agility Robotics, Apptronics and Reflex Robotics in its warehouses. These robots perform tasks such as moving containers and recycling materials. Although the technology is still in the early testing phase, it could revolutionize efficiency in the logistics industry.

Google DeepMind introduces Gemini Robotics

Google DeepMind has introduced two new AI models, Gemini Robotics and Gemini Robotics-ER, based on the Gemini 2.0 model. These models enable robots to perform complex physical tasks and adapt to new environments. This marks a significant advance in the development of versatile and intelligent robots for real-world applications.

Zyphra Launches ZR1-1.5B: A Compact Powerhouse for Math and Code

Zyphra has announced the release of its first reasoning model, ZR1-1.5B. Despite its compact size, the model delivers impressive performance in both math and coding tasks, ranking among the top in its category. ZR1-1.5B is also highly efficient, using 60% fewer reasoning tokens compared to similar models. This makes it a strong contender for lightweight, high-performance AI applications.