Tencent: Hunyuan A13B Instruct

Tencent's Hunyuan A13B Instruct model, a 13B active parameter MoE model with 80B total parameters, demonstrates exceptional performance across several key metrics. It consistently ranks among the fastest models, achieving an Infinityth percentile across six benchmarks, and offers highly competitive pricing, also at an Infinityth percentile across five benchmarks. The model exhibits outstanding reliability with a 100% success rate across all evaluated benchmarks, indicating minimal technical failures. While excelling in speed, cost, and reliability, the model's benchmark performance presents a mixed picture. It shows strong capabilities in Reasoning and General Knowledge, achieving 94.0% and 97.5% accuracy respectively, with its Reasoning performance being particularly notable as the most accurate model at its price point. Coding also demonstrates solid performance at 90.0% accuracy. However, the model exhibits a critical weakness in Instruction Following, scoring 0.0% accuracy in both evaluated instances, suggesting a significant limitation in processing complex or multi-layered directives. Its Email Classification accuracy is also comparatively lower at 91.0% (20th percentile). Overall, Hunyuan A13B Instruct is a highly efficient and cost-effective model, particularly strong in reasoning and general knowledge, but requires significant improvement in instruction following.