The Curious Case of ChatGPT API Pricing

The Curious Case of ChatGPT API Pricing

Aman Jha, the founder of?Explainpaper,?said?that their runway just doubled – ‘at least’. “We were right to keep Explainpaper free for so long,” he added, saying that it was possible because of the 10% cost benefits offered by the all-new?ChatGPT API.

ChatGPT has already taken the?internet by storm, and now its API is making waves in the enterprise and developer ecosystems. Another interesting development, along with the API launch, was the?reduction in cost. All of this will now increase the user adoption rate.?

But, one question remains on how they were able to reduce the cost. Simply put, through a series of system-wide optimisations,?OpenAI?achieved a?90 percent cost reduction?for ChatGPT?since December, and they are now passing the cost benefits to API users.?

Check out the price comparison chart below:

No alt text provided for this image

Assimilation Mode

This is the?second time in a year?that OpenAI has managed to drop the price. But why? Data science expert and founder of V Squared,?Vin Vashishta?pointing at – ‘competition and adoption,’?said?that the move will drive early-stage startup competitors out of the market and make it too expensive for new entrants. “Lower pricing makes their competitors' path to profitability much harder,” he added, saying that lower margins leave?competitors?with less money to fuel innovation and model improvement cycles.?

Further, Vashishta said that it hurts margins, but they are making it up with scale. “Cheaper to use means easier to adapt too. It is an intelligent path to rapidly gaining market share,” he added. He said that OpenAI is using model development and?training costs?as a competitive advantage. Their best-in-class large model optimisation capabilities mean they can preserve margins while less experienced and knowledgeable?competitors struggle.?

While profitability is critical to driving the business, what about the carbon emission concerns arising from hundreds and thousands of developers using ChatGPT API??Read to find out.


AIM Research >>

Utilisation of Data Pipeline Architecture by Indian Enterprises

It's critical for organizations to have a defined data pipeline architecture.

No alt text provided for this image

???? We recently did a study, in collaboration with?Hansa Cequity, that highlights the challenges faced by mid to large-scale Indian companies in building effective data pipelines. The study finds that only 37% of organizations have set methodologies and processes in place to measure the quality of data ingested in the pipeline.

  • 71% of companies feel that they are not deriving optimal value from the data at their disposal due to various factors such as time constraints, lack of quality data, high cost, and talent shortage.

?? However, investing in a robust data pipeline architecture can help organizations overcome these challenges and enable real-time analytics for faster and more effective data-driven decision-making.

?? To learn more about the current state of data pipeline architecture in Indian enterprises, visit this?link.


Top Stories of the Week >>

Hugging Face Makes OpenAI’s Worst Nightmare Come True

Last week, EleutherAI announced its plans to become a?nonprofit foundation?called the ‘EleutherAI Institute’. Founded by a?decentralised group of researchers, the entity has unsurprisingly been backed by?Hugging Face?and Emad Mostaque’s Stability AI, former GitHub CEO Nat Friedman, Lambda Labs and Canva. All of these companies are technically still ‘independent’ in their business segments.?

A few week’s before that Hugging Face?partnered with AWS?with the scope to integrate their models into the AWS cloud. There seems to be a joint effort by companies to promote the open-source ecosystem. More precisely, they are making OpenAI’s worst nightmare come to life.?Read more here.

Killing Prompt Engineering?

Prompt engineering has become a hot topic ever since the popularity of ChatGPT and other large language models surfaced. It has also been making a surge in courses, resource materials, job listings and whatnot. However, experts believe with models getting better, the need for prompt engineering is slowly dying.?Read more here.?

When ChatGPT Attempted UPSC Exam

This story has?officially gone viral. We at?AIM?have been experimenting with ChatGPT for a while now and were curious to see if it was able to crack tough exams and interviews. In that spirit, we wondered, ‘Can ChatGPT clear UPSC?’?Read to find out what happened next.


People & Tech >>

Meet Mira Murati, the 35-year-old CTO of OpenAI

A few weeks back, the spotlight shifted to?Mira Murati, the CTO of AI startup OpenAI. While many Indian media outlets?claimed her origin as Indian, Murati is actually of Albanian descent. Born and raised in San Francisco to high school teachers from Vlor?, Murati's sudden emergence from obscurity has piqued the interest of those following?OpenAI.?Read more here.


AIM Videos >>

AIM?recently got in touch with?Arkadiy Dobkin, the founder and CEO of EPAM Systems, where we touched upon various aspects, including the demand side of things for ChatGPT, the large language model as a service and more.?


AIM Shots >>

Amid the rising privacy concerns, Google?recently announced?the improvement of the model’s privacy by combining secure aggregation (SecAgg) and a distributed version of?differential privacy.?

AI robotics startup?Figure?recently announced the launch of the?world’s first commercially available general-purpose humanoid robot?Figure O1, the prototype of which bears a close resemblance to Tesla’s?Optimus.

A team of researchers at Google and other institutions have introduced the?Universal Speech Model?(USM) – a single large model that performs automatic speech recognition (ASR) in?more than 100 languages.

A group of creators at the VFX company Corridor Digital recently shared a fantastic anime video – made on Stable Diffusion.?Artists are not happy.?Read to find out.?

NVIDIA rival?Sapeon?recently raised funding, taking its valuation well above $400 million.?Check out the details here.?

Free and open-source network packet analyzer software?Wireshark?last week announced its plan to turn into an?open-source foundation.?Read more here.?

Meta recently announced its?AR/VR roadmap. This time it is for real.?Check out the details here.

An engineer from?Replicate, who goes by the name ‘zeke’ on GitHub has created?Scribble Diffusion, an open-source app that can convert your hand-drawn art, along with a text prompt, into an?art piece.?Try out the app here.

Applied Materials?recently unveiled a breakthrough?lithography technology?which reduces the cost, complexity, and environmental impact of?advanced chipmaking?significantly.?

Google recently integrated?Google Dataset?into its Google search query.

Google has released?Dreamix, an AI model that performs?text-based motion and appearance editing?of general videos.?Read more here.?

Microsoft recently released its research paper,?Language Is Not All You Need: Aligning Perception with Language Models. The model introduces a?multimodal large language model?(MLLM) called?Kosmos-1.

Tesla chief?Elon Musk?has been approaching AI researchers to set up his own research lab to develop an alternative to ChatGPT.?Read more here.

Raj Bhatt

CEO @ Knowledge Foundry | Analytics, Machine Learning

1 年

Two things that could be driving the pricing: 1) While the turbo API is now available, there are no SLAs on availability or turnaround time. OpenAI down time has become a serious issue. 2) OpenAI will get everyone hooked onto turbo before Google ups its LLM API game.

要查看或添加评论,请登录

社区洞察

其他会员也浏览了