Welcome to the forefront of conversational AI as we explore the fascinating world of AI chatbots in our dedicated blog series. Discover the latest advancements, applications, and strategies that propel the evolution of chatbot technology. From enhancing customer interactions to streamlining business processes, these articles delve into the innovative ways artificial intelligence is shaping the landscape of automated conversational agents. Whether you’re a business owner, developer, or simply intrigued by the future of interactive technology, join us on this journey to unravel the transformative power and endless possibilities of AI chatbots.
National
ETV Bharat / technology
By ETV Bharat Tech Team
Published : September 21, 2026 at 11:40 AM IST
Hyderabad: A new Artificial Intelligence (AI) startup called TypeSafe AI has launched its own AI model named Jev, joining the ranks of well-known chatbots such as OpenAI’s ChatGPT, Anthropic’s Claude, and Google’s Gemini.
The new AI model was launched on September 15, 2026, led by Diogo Almeida, a former OpenAI researcher who helped build ChatGPT.
At OpenAI, Almeida was a co-creator of Reinforcement Learning from Human Feedback (RLHF), a method used to train ChatGPT. Despite AI’s growing power, Almeida felt current models were not very effective for automation tasks. This dissatisfaction led him to leave OpenAI and start TypeSafe AI to solve the problem.
After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI?
I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev
• 20-200x faster
• 40-400x… pic.twitter.com/JSybNG2BKJ
A different kind of AI model
TypeSafe AI developed Jev under the company’s first System One Model, which is said to be a new class of frontier AI models built to make fast, structured decisions for software automation.
Unlike typical AI chatbots, Jev is not a standard large language model (LLM). Instead of generating text, it gives what the company calls “calibrated decisions” based on set probabilities.
As Jev’s answers come from a fixed set of outputs, the company claims it makes fewer mistakes, works faster, and costs less to run.
Pricing
Jev costs $0.042 per million input tokens and free output tokens, as TypeSafe AI says it’s “too cheap to metre”.
How Jev works?
When a user types a query on their automated software, Jev receives two things: the State and the Typed Question.
The State is the raw background information like a customer email, financial data, medical log, etc., while the Typed Question is the specific thing that the user wants Jev to evaluate based on the State. The AI chatbot, instead of reading the background information as text and thinking about it conversationally, Jev analyses the information to answer users’ questions simultaneously. This is because Jev does not generate a sentence word by word like a typical chatbot; instead, it scans the data and calculates the answer in a quick, single go.
Jev entirely skips the conversational text and puts the output in clean, structured data points, which are shown in three ways: Noul, Choice, and Score.
TypeSafe AI says it trained the System One and Jev using a technique called Reinforcement Learning for Calibrated Decisions. Using this technique, the company says the AI model’s confidence level matches its actual accuracy. This means that when confidence is high, the system can act on its own, but when confidence is low, it can pass the decision to a human instead.
Developers have already shown strong interest in Jev for software automation. Demand has been so high that TypeSafe AI’s API was temporarily closed to new users.
Companies such as Vercel, Cloudflare, LangChain, and Langfuse added Jev to their stacks within three days of its launch.
Praneet Sharma, a software engineer at Vercel, said his company had earlier used an OpenAI model to run a command safety classifier. After switching to Jev, he said results came back 5 to 18 times faster, with better accuracy.
We benchmarked fx auto mode (safety) classifier with @typesafeai‘s Jev.
tl;dr: ~5-18x faster and more accurate than 𝚐𝚙𝚝-𝟻.𝟼-𝚕𝚞𝚗𝚊, our current top choice pic.twitter.com/3G8tpRz7AG
Named after an economist
TypeSafe AI says the model is named after economist William Stanley Jevons, known for the Jevons Paradox. This theory suggests that when a resource becomes cheaper, people often use more of it rather than less. The company believes the same pattern could apply to AI as it becomes more efficient and widely used.
For All Latest Updates
TAGGED:
Copyright © 2026 Ushodaya Enterprises Pvt. Ltd., All Rights Reserved.