PyyanThe AI Index

Pyyan / Compare / DeepSeek V4 Flash Vision Exp vs Inkling

DeepSeek V4 Flash Vision Exp vs Inkling

2 of 5

Open-Weight Models · verified 13 Sept 2026

DeepSeek V4 Flash Vision ExpDeepSeekdeprecated
TInklingThinking Machines Labcurrent
3 slots left
SpecificationDeepSeek V4 Flash Vision ExpInkling
SummaryExperimental, says so in the name, and shipped anyway.975B under Apache 2.0, shipped as a starting point rather than a finished product.
Context128K1M
Input ($/Mtok)$0.44Weights only
Output ($/Mtok)$1.32Weights only
LicenceOpen weightsApache 2.0
Max output32KNot published
ParametersNot published975B total, 41B active
Released21 Aug 202615 Jul 2026
Model IDdeepseek-ai/DeepSeek-V4-Flash-Vision-ExpNot published
CategoryOpen-Weight ModelsOpen-Weight Models
Official—Thinking Machines Lab ↗

Highlighted rows are where these differ.

DeepSeek V4 Flash Vision Exp

  • Retired 10 September 2026: requests now route to DeepSeek V4.1 Flash
  • Was DeepSeek's first Flash tier with image input

Best for cheap multimodal trials.

Full spec sheet →

Inkling

  • Pretrained on 45 trillion tokens of text, images, audio and video, with an encoder free architecture
  • Apache 2.0, weights on Hugging Face, same day fine tuning through the lab's Tinker platform
  • The lab earns from the compute customers use to adapt it, not from per token access

Best for fine tuning on your own data.

Full spec sheet →