AI Hub
← All models

Llama 3.2 3B Instruct vs Sonar Reasoning Pro

Meta vs Perplexity — benchmarks, pricing, and capabilities side by side.

  • Sonar Reasoning Pro has the higher intelligence index (95.7 vs 30.8)
  • Llama 3.2 3B Instruct is cheaper ($0.05 vs $2.00 per 1M input)
  • Llama 3.2 3B Instruct has a larger context window (131K)
Llama 3.2 3B InstructSonar Reasoning Pro
Intelligence index30.895.7
DeveloperMetaPerplexity
TypeLLMMultimodal
AccessOpen weightsAPI only
Context window131,072 tokens128,000 tokens
Input price$0.05 / 1M$2.00 / 1M
Output price$0.34 / 1M$8.00 / 1M
Speed172 tok/s
ReleasedSeptember 25, 2024March 7, 2025
Parameters
Input modalitiesTextText, Image
Output modalitiesTextText

Shared benchmarks

Llama 3.2 3B Instruct
Sonar Reasoning Pro
MATH-500
48.9
95.7