AI Hub
← All models

DeepHermes 3 - Llama-3.1 8B vs DeepSeek-V4-Flash

Nous Research vs DeepSeek — benchmarks, pricing, and capabilities side by side.

  • DeepSeek-V4-Flash has the higher intelligence index (89.4 vs 23.5)
  • DeepHermes 3 - Llama-3.1 8B is cheaper ($0.00 vs $0.10 per 1M input)
DeepHermes 3 - Llama-3.1 8BDeepSeek-V4-Flash
Intelligence index23.589.4
DeveloperNous ResearchDeepSeek
TypeLLMLLM
AccessOpen weights
Context window1,048,576 tokens
Input price$0.00 / 1M$0.10 / 1M
Output price$0.00 / 1M$0.20 / 1M
Speed109 tok/s
ReleasedFebruary 13, 2025April 24, 2026
Parameters284B (13B active)
Input modalitiesText
Output modalitiesText

Shared benchmarks

DeepHermes 3 - Llama-3.1 8B
DeepSeek-V4-Flash
GPQA Diamond
27
89.4
Humanity’s Last Exam
4.3
32.1