Pyyan / Compare / Inkling 2 Small vs Qwen3.8-Flash-Next

Inkling 2 Small vs Qwen3.8-Flash-Next

2 of 5

Open-Weight Models · verified 14 Nov 2026

TInkling 2 SmallThinking Machines Labcurrent
Qwen3.8-Flash-NextAlibabacurrent
3 slots left
SpecificationInkling 2 SmallQwen3.8-Flash-Next
SummaryA 594B multimodal model with 17B active, trained natively in 4-bit NVFP4, under Apache 2.0.The most downloaded thing on Hugging Face right now, and a stated Qwen4 preview.
Context1M262K native, ~1M extended
Input ($/Mtok)Weights only$0.15
Output ($/Mtok)Weights only$0.47
LicenceApache 2.0Open weights
Max outputNot published32K
Parameters594B total, 17B active125B total, 6B active
Released5 Nov 2026, weights 12 Nov26 Aug 2026
Model IDthinkingmachines/Inkling-2-SmallQwen/Qwen3.8-Flash-Next
CategoryOpen-Weight ModelsOpen-Weight Models
OfficialThinking Machines Lab ↗—

Highlighted rows are where these differ.

Inkling 2 Small

  • Weights published on Hugging Face on 12 November 2026, a week after a hosted-only launch on Tinker on 5 November
  • Takes text and images, writes text, with a context window of up to 1M tokens
  • Pretrained, midtrained and post-trained in NVFP4. The main checkpoint needs at least 500GB of GPU memory, such as 2 B300s or 4 H200s
  • An MXFP8 checkpoint is also published and needs at least 800GB. The lab says it scores only 1 to 2 points higher
  • Local serving recipes for SGLang, vLLM, TokenSpeed and Unsloth. The benchmark scores on the card are the lab's own

Best for self-hosting a large text and image model on a multi-GPU node, or fine-tuning it on Tinker.

Full spec sheet →

Qwen3.8-Flash-Next

  • 125B total parameters with only 6B active per token
  • Alibaba describes it as a preview of the Qwen4 architecture
  • Top of the Hugging Face trending list, with a GGUF build close behind it

Best for open weights at speed.

Full spec sheet →