qwen2.5:0.5b
Upstream reference: Qwen2.5-0.5B-Instruct
Publisher: Qwen Team
Qwen presents this checkpoint as the instruction-tuned 0.5B member of Qwen2.5, using a causal language-model architecture for text interaction.
Publisher-described context, separate from Sigma5C observations and recorded catalog statistics.
Published upstream specifications
- Parameters
- 0.49 billion Source
- Published context
- 32,768 tokens Source
- Published generation limit
- 8,192 tokens Source
- Layers
- 24 Source
- Upstream license label
- Apache-2.0 Source
Publisher-described uses
- Instruction-following text generation. Source
Reference limitations
- The exact checkpoint card gives 32,768 tokens; the family-level 128K statement must not replace that checkpoint-specific value. Source
A recorded tag and an upstream card association do not verify the deployed checkpoint, quantization, configured context or current hosted availability.
Publisher sources
- Qwen2.5-0.5B-Instruct model card — Qwen Team. Checked .
Links
Recorded references do not establish downloadable weights, licensing or current availability.
0.5B
parameters
Q4_K_M
quantization
qwen2
architecture
32768
context
apache-2.0
license
2024-09-19
released
Where qwen2.5:0.5b comes from
Company / lab
Alibaba Cloud Qwen
See every model and family from this maker.
Model family
qwen2.5
Other versions and sizes in this line.
completiontools
Named in 7 observation pages
- A consumer GPU recorded higher throughput than a 48-core Xeon · V1
- How models scale under concurrent slots · V1
- NVIDIA GTX 1070 — 344 models profiled · V1
- Sub-3B local models — all 34 · V1
- The capability ladder — execution-verified · V1
- The minimum viable model probe · V1
- Xeon vs AMD workstation — CPU inference head-to-head · V1