Microsoft integrates OpenAI and Anthropic models in its AI agent Copilot to offer more precise answers. This Monday, from New York, the company launched Critique, a function that simultaneously consults ChatGPT of OpenAI and Claude of Anthropic in each query.
The revolution of Critique in Copilot
Before, Copilot relied on a single language model, which limited its depth. Now, Microsoft integrates OpenAI and Anthropic models to compare results and select the most reliable option. This reduces errors and improves utility in complex tasks such as academic research or business analysis.
How does this integration work?
- The user sends a query to the Copilot research agent.
- Critique processes the question with both OpenAI and Anthropic models.
- Generates a synthesized response, highlighting strengths of each one.
According to AI experts, this multi-model approach elevates precision up to 20-30% in independent benchmarks, based on tests with public datasets such as MMLU.
Background of the collaboration
Microsoft integrates OpenAI and Anthropic models after years of strategic alliances. OpenAI, with whom Microsoft invested billions since 2019, provides GPT-4. Anthropic, backed by Amazon, offers Claude 3.5 Sonnet, known for its security. This move responds to fierce competition: Google with Gemini and Meta with Llama.
Impact on users and the Dominican market
In the Dominican Republic, where access to AI tools grows 40% annually according to local digital adoption data, this novelty benefits students, entrepreneurs, and professionals. Copilot, integrated in Microsoft 365, now directly competes with unipolar solutions. Analysts predict that it will drive global productivity, with projections of saving 1 billion hours of labor for 2026.
Key advantages of combined models
ChatGPT stands out in creativity; Claude, in ethical reasoning. Together, they minimize common hallucinations in AI. Microsoft plans to expand Critique to more agents, such as in Azure.
This strategy positions Microsoft as a leader in hybrid AI. Microsoft integrates OpenAI and Anthropic models not only maximizes precision but redefines standards in accessible technology for all.
