AI Hub
← All models

DeepSeek-V4-Flash vs Llama 3 8B Instruct

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

  • DeepSeek-V4-Flash has the higher intelligence index (89.4 vs 32.4)
  • Llama 3 8B Instruct is cheaper ($0.04 vs $0.10 per 1M input)
  • DeepSeek-V4-Flash is faster
  • DeepSeek-V4-Flash has a larger context window (1M)
DeepSeek-V4-FlashLlama 3 8B Instruct
Intelligence index89.432.4
DeveloperDeepSeekMeta
TypeLLMLLM
AccessOpen weightsOpen weights
Context window1,048,576 tokens8,192 tokens
Input price$0.10 / 1M$0.04 / 1M
Output price$0.20 / 1M$0.04 / 1M
Speed109 tok/s81 tok/s
ReleasedApril 24, 2026April 18, 2024
Parameters284B (13B active)
Input modalitiesTextText
Output modalitiesTextText

Shared benchmarks

DeepSeek-V4-Flash
Llama 3 8B Instruct
GPQA Diamond
89.4
29.6
Humanity’s Last Exam
32.1
5.1