AI Hub
← All models

DeepSeek-V4-Flash vs Jamba Large 1.7

DeepSeek vs AI21 Labs — benchmarks, pricing, and capabilities side by side.

  • DeepSeek-V4-Flash has the higher intelligence index (89.4 vs 36.5)
  • DeepSeek-V4-Flash is cheaper ($0.10 vs $2.00 per 1M input)
  • DeepSeek-V4-Flash is faster
  • DeepSeek-V4-Flash has a larger context window (1M)
DeepSeek-V4-FlashJamba Large 1.7
Intelligence index89.436.5
DeveloperDeepSeekAI21 Labs
TypeLLMLLM
AccessOpen weightsOpen weights
Context window1,048,576 tokens256,000 tokens
Input price$0.10 / 1M$2.00 / 1M
Output price$0.20 / 1M$8.00 / 1M
Speed109 tok/s48 tok/s
ReleasedApril 24, 2026August 8, 2025
Parameters284B (13B active)
Input modalitiesTextText
Output modalitiesTextText

Shared benchmarks

DeepSeek-V4-Flash
Jamba Large 1.7
GPQA Diamond
89.4
39
Humanity’s Last Exam
32.1
3.8