OpenAI has lately responded to the threat posed by Chinese DeepSeek AI chatbot with the early release of o3-mini and aims to defend its position in the AI market. The R1 model of DeepSeek made waves in recent weeks by delivering top-tier performance amid consuming less computational power. OpenAI announced that its newest o3-mini model in its reasoning series is now available on both ChatGPT and the API. The company has also introduced free reasoning capabilities for users as well as increased message limits for paid subscribers.
The “o” models focus on reasoning and structured problem-solving. The o3-mini comes in low, medium and high variants to provide a range of performance levels suited for various tasks.
The announcement of R1 caused a massive selloff in the tech sector earlier and Nvidia losing $600 billion in market value as investors reassessed the demand for high-cost AI chips. DeepSeek has optimized the way AI processes information and this has intensified competition.
OpenAI has made the o3-mini model 24% faster than its predecessor and is also learned to have improved cost-effectiveness. The pricing for o3-mini is competitive, but still remains higher than the R1 rates of DeepSeek. OpenAI’s model is closed-source while DeepSeek R1 is freely available.
Benchmark tests reveal that o3-mini performs well against DeepSeek R1 and particularly at higher tiers. The o3-mini medium scores slightly lower than R1 in some tests whereas the high-end version outperforms it in math and coding tasks. The overall performance of both models is comparable and the choice between them may depend on specific use cases.
OpenAI o3-mini struggled with multi-step reasoning tasks compared to DeepSeek R1 in practical evaluations. DeepSeek R1 correctly identified the answer in one test involving a detective-style puzzle whereas o3-mini provided an incorrect response.
The AI race is gradually intensifying at a rapid pace equipped with OpenAI and DeepSeek pushing each other to innovate. DeepSeek has disrupted the industry with its efficiency-driven approach while OpenAI remains a dominant player with its advanced reasoning capabilities.