OpenGVLab: InternVL3 2B

The OpenGVLab: InternVL3 2B model, created on April 30, 2025, is positioned as a multimodal large language model emphasizing speed and reasonable performance. It performs among the fastest models, typically ranking in the top tier for speed (70th percentile across benchmarks). Furthermore, it consistently offers highly competitive pricing, placing it in the 88th percentile. In terms of performance across benchmark categories, the InternVL3 2B demonstrates significant weaknesses in knowledge-based tasks. Its accuracy on General Knowledge is notably low at 15.0% (12th percentile), and similarly, its Ethics performance is 27.0% (12th percentile). While these benchmarks show low accuracy, the model maintains competitive costs and moderate durations for these tasks. Its performance in the Email Classification task is also low at 79.0% accuracy (8th percentile), despite a very competitive cost and a relatively fast duration. Overall, the model's primary strength lies in its speed and cost-efficiency, making it an attractive option for applications where these factors are paramount, even if it comes at the expense of accuracy in complex reasoning and classification tasks.