Today : Jan 31, 2025
Technology
31 January 2025

Microsoft Makes OpenAI's O1 Reasoning Model Free For Copilot Users

New 'Think Deeper' feature aims to boost AI accessibility and support complex queries.

Microsoft has made headlines this week by announcing the availability of OpenAI's o1 reasoning model for free to all users of its Copilot service, via the new "Think Deeper" feature. The announcement came from Mustafa Suleyman, CEO of Microsoft AI, who confirmed through a LinkedIn post, "Think Deeper is now free and available for all users of Copilot." This significant move allows users to access advanced AI capabilities typically locked behind subscription fees.

Previously, the o1 reasoning model was available only to paying subscribers of ChatGPT Pro and ChatGPT Plus, costing $200 and $20 monthly respectively. Now, users can simply click the "Think Deeper" button on their Copilot interface, which enhances the AI's capacity to tackle complex queries by taking additional time—around 30 seconds—to generate responses. This time extension helps the model process information more thoughtfully and is particularly useful for handling challenging topics.

Essentially, the o1 model operates like someone showing their work on mathematical problems: it provides insights and steps taken to arrive at conclusions. This contrasts with traditional AI models, which often generate responses almost instantaneously. OpenAI has stated, "o1 uses a chain of thought when attempting to solve a problem. Through reinforcement learning, o1 learns to hone its chain of thought and refine the strategies it uses." This learner model stands out for its unique approach to AI reasoning.

The timing of Microsoft’s announcement closely aligns with the recent rise of DeepSeek's Chinese AI app, which also gained popularity by providing multi-step reasoning answers to users for free. Microsoft’s move appears strategically calculated, likely intended to attract curious users who might otherwise be drawn to alternatives like DeepSeek's offerings. While DeepSeek faces scrutiny over severe privacy issues—exposing chat logs for anyone to see—Microsoft seems to position itself as the more reliable choice for informed AI interaction.

There’s no doubt about the significance of this shift. Not only does it create more accessibility for users, but it also reinforces Microsoft’s and OpenAI's joint effort to lead the AI conversation at no cost to the consumer. The actual rollout of the "Think Deeper" feature might still be gradual; users are encouraged to check back if they do not see it available immediately.

Users who explore the new feature will find it appropriate for research purposes, helping to deepen their insights on various subjects through well-rounded explanations. Given the o1 model's strengths, such as reportedly being able to solve 83% of International Mathematics Olympiad problems compared to only 13% by its predecessor GPT-4, there’s reason to expect impressive results.

It's important to note, though, even with these advancements, users shouldn’t assume 100% accuracy from the responses generated. The generative nature of these AI outputs means they can still produce errors or hallucinations, even on complex topics. Therefore, Microsoft advises caution: reliance on these answers should be measured and complemented by independent verification.

Accessing the o1 reasoning model will require users to sign in with their Microsoft accounts but won’t incur any charges, following Microsoft’s recent trend toward democratizing advanced AI tools across its platforms. The Copilot application can be accessed through either standalone Windows apps or directly through copilot.microsoft.com.

Looking forward, this free availability raises intriguing questions about the future of AI technology. With deepening competition among major players, including notable offerings from Google and the rise of specialized services like DeepSeek, how these tools evolve will largely depend on user experience and accessibility.

It’s also significant to acknowledge the broader environment: AI applications are rapidly developing. Recent announcements hint at newer iterations, like OpenAI’s forthcoming o3 model, which promises even more advanced reasoning capabilities through what they call "private chain of thought." Though it seems unlikely this will be offered for free, it’s part of the fast-paced world where innovation is continuously ushering change.