smallthinker:3b

Upstream reference: SmallThinker 3B Preview

Publisher: Tiiny / PowerInfer

The publisher documents a Qwen2.5-3B-Instruct derivative fine-tuned using long reasoning traces, intended for constrained devices and draft-model research.

Publisher-described context, separate from Sigma5C observations and recorded catalog statistics.

Published upstream specifications

Base lineage
Qwen2.5-3B-Instruct Source
Documented training style
Supervised fine-tuning on long reasoning-trace datasets Source

Publisher-described uses

  • Small reasoning-model and speculative draft-model research. Source

Reference limitations

  • The card notes English-only training, instruction-following limits, repetition and unexpected outputs. The original PowerInfer card redirects to Tiiny; retained historical attribution should not be silently rewritten. Source

A recorded tag and an upstream card association do not verify the deployed checkpoint, quantization, configured context or current hosted availability.

Publisher sources

Links

Recorded references do not establish downloadable weights, licensing or current availability.

3.0B
parameters
Q8_0
quantization
qwen2
architecture
32768
context
2024-12-12
released

Where smallthinker:3b comes from

Company / lab

PowerInfer / Tiiny (publisher)

See every model and family from this maker.

Model family

smallthinker

Other versions and sizes in this line.

completionthinking

Named in 1 observation page

Browse the Observatory

AI-generated & AI-scored. Every observation on this site is produced and judged by AI, and AI makes mistakes. Treat each result as a recorded observation under a declared protocol, as of its date. Method and marker legend