Join our daily and weekday newsletters to receive the latest updates on AI coverage. Learn More
Openai has announced that it will be releasing its powerful OpenAI software today. Deep Research is available to all ChatGPT Plus (#19459103) Team Education Enterprise users will now have access to the AI agent that many experts believe is the most revolutionary AI agent since ChatGPT. OpenAI has announced that
according to an announcement. Official X account Plus, Team, Education and Enterprise will receive 10 deep-research queries per month. Pro tier subscribers will be able to access 120 queries each month.
Deep Research is powered by an advanced version of OpenAI’s upcoming o3 modelrepresents a significant change in how AI can help with complex research tasks. Deep Research, unlike traditional chatbots, which provide immediate responses, independently searches hundreds of online sources. It analyzes text, PDFs, images, and other formats, and synthesizes reports that are comparable to those produced professionally.
Deep research is now available to all ChatGPT Plus users, including Team, Edu and Enterprise users.
– OPENAI (@OpenAI).””https://twitter.com/OpenAI/status/1894454194943529433?ref_src=twsrc%5Etfw””> February 25, 2025 (19659010)
The AI arms race: DeepSeek’s open challenge meets OpenAI’s premium play
It is not a coincidence that OpenAI has expanded its rollout. The generative AI landscape changed dramatically in the last few weeks, with China’s DeepSeekis a disruptive force. By open-sourcing their DeepSeek R1 model is available under an My License – the company has fundamentally questioned the closed, subscription based business model which has defined Western AI development.
The divergent philosophies are what makes this competition so interesting. While OpenAI continues enclosing its most powerful capabilities behind an increasingly complex DeepSeek’s subscription tiershave chosen a radical new approach: give away the technology, and let a million applications blossom.
Chinese artificial intelligence company Deepseek made waves recently when it announced R1, a reasoning model open-source that it claimed was comparable to OpenAI’s O1 at a fraction the cost.
The strategy is reminiscent of earlier eras in technology adoption when open platforms created more value. Linux’s dominance of server infrastructure is a compelling historical comparison. Enterprise decision-makers must decide whether to invest in proprietary software that may provide immediate competitive advantages, or open alternatives that can foster broader innovation within their organization.
Perplexity’s The recent integrationDeepSeek R1 into its own tool at a fraction the price of OpenAI’s shows how quickly an open approach can produce competing products. Anthropic Claude 3.7 Sonnet, on the other hand, has taken a different path, focusing transparency in its reasoning with “visible extended thought.”
The r1 from deepseek is a very impressive model. Especially in terms of what they can deliver for the price.
We will deliver much better models, and it’s also legitimatingly invigorating to be a new rival! We will pull some releases.
– Sam Altman (@sama).””https://twitter.com/sama/status/1884066337103962416?ref_src=twsrc%5Etfw””> January 28, 2025
This fragmentation has led to a market where every major player offers a unique approach to AI-powered Research. This means more choice for enterprises, but also greater complexity in determining the platform that best aligns with specific needs and values.
OpenAI’s calculated democratic pivot from walled garden into public square
Sam Altman’s statement that Deep Research is “probably worth $1,000 a monthly to some users” reveals more than price elasticity. He acknowledges the extraordinary value disparity among potential users. This admission is at the core of OpenAI’s ongoing strategy balancing act.
OpenAI faces a fundamental conflict: maintaining the exclusivity of its premium products that fund its development, while also fulfilling its mission to ensure that “artificial intelligence benefits all humanity”. Today’s announcement represents an effort toward greater accessibility without undermining their revenue model.
I think we will initially offer 10 uses for chatgpt+ and 2 per monthly in the free tier. We plan to increase these over time.
It’s probably worth $1000 per month to some users, but i am excited to see what everybody does with it. https://t.co/YBICvzodPF
– Alone Altman (@sama)””https://twitter.com/sama/status/1889679679696806103?ref_src=twsrc%5Etfw””> February 12, 2025
OpenAI’s free tier will limit users to two queries per month, which is enough to demonstrate its capabilities without compromising the premium offerings. This approach follows the “freemium playbook” that has defined the digital economy. However, it comes with some unusually strict constraints to reflect the significant computing resources required for every Deep Research query.
Allocation of This clear distinction between Pro users ($200/month), who get 120 queries per month, and Plus users (20/month), creates a clear value proposition. OpenAI’s tiered rollout strategy indicates that it recognizes the need to democratize access to advanced AI capabilities. This requires more than simply lowering price barriers. It also necessitates fundamentally rethinking how these capabilities are packaged.
Headline figure — 26.6% accuracy in ” Humanity’s last exam ” — tells just part of the story. This benchmark is designed to be extremely challenging, even for human experts. It represents a quantum jump beyond previous AI capabilities. Even achieving 10% on this test was considered impressive just a year before.
The test’s nature is more important than the raw performance. It requires a combination of information from disparate domains, and nuanced reasoning beyond pattern matching. Deep Research’s method combines several technological advances: multi-stage planning and adaptive information retrieval, and, perhaps more importantly, a form self-correcting algorithm that allows the system to identify and correct its own limitations throughout the research process. These capabilities are not without their blind spots. The system is still vulnerable to what could be called ” Consensus biasis a tendency to favor widely accepted viewpoints, while potentially overlooking counter-arguments that challenge established thinking. This bias is particularly problematic in areas where innovation often comes from challenging conventional wisdom.
The system inherits the biases, and limitations of the source material because it relies on existing web content. Deep Research may struggle in rapidly evolving fields, or niche specialties that have limited online documentation. Without access to academic journals or proprietary databases, Deep Research’s insights into certain specialized fields may be superficial, despite its sophisticated reasoning abilities.
OpenAI’s Deep Research tool outperforms competitors on Perplexity Labs’ “Humanity’s Last Exam” benchmark, achieving roughly 25% accuracy — significantly ahead of other AI models including those from Perplexity, DeepSeek, Google and Anthropic. (Credit: Perplexity Labs)
The Executive e’s quandary: How Deep Research redefines the rules of knowledge-based work
Deep Research presents C-suite executives with a paradox. It’s powerful enough to redefine roles within their organization, yet it’s still too limited to deploy without careful human supervision. The immediate productivity gains of Deep Research are undeniable. Tasks that used to take days of analyst work can now be completed within minutes. This efficiency has complex strategic implications.
Organizations who integrate Deep Research effectively may need to completely reimagine their workflows. Instead of replacing junior analysts, this technology could create new hybrid roles, where human expertise is focused on framing questions and evaluating sources, as well as critically assessing AI generated insights. Deep Research will be viewed as a tool that enhances human capabilities, not as a substitute for human judgment.
Deep research for chatgpt Plus users!
One of my favorite things that we have ever shipped.
– Alone Altman, @sama””https://twitter.com/sama/status/1894527988378550392?ref_src=twsrc%5Etfw””> The pricing structure has its own strategic considerations. Each query costs $1.67 at $200 per month for Pro users who have 120 queries. This is a small cost compared to the costs of human labor. The limited volume forces organizations to prioritize questions that truly deserve Deep Research’s capabilities. Ironically, this constraint could lead to a more thoughtful application of technology than would be possible with a pure unlimited model.
Long-term implications will be more profound. As research capabilities, which were previously only available to elite organizations, become more widely available, competitive advantage will come from the way organizations frame their questions and integrate AI generated insights into their decision making processes. The strategic value shifts away from information gathering and towards insight generation.
The message for technical leaders is clear: the AI research revolution has arrived. The question isn’t whether to adapt, but how quickly can organizations develop the skills, processes and cultural mindset required to thrive in an environment where deep research has fundamentally been democratized.
VB Daily provides daily insights on business use-cases
Want to impress your boss? VB Daily can help. We provide you with the inside scoop about what companies are doing to maximize ROI, from regulatory changes to practical deployments.
Read our privacy policy
Thank you for subscribing. Click here to view more VB Newsletters.
