Unverified50% confidenceBenchmarkExact time
In testing, all three models—qwen2.5:3b, qwen2.5:1.5b, and llama3.2:1b—kept model memory below 0.5 GB, and all produced 0 hallucinated place names.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/11/2026
First Seen
Valid until: 1/9/2027
Sources
Related Entities
Related Claims
UnverifiedThe study tested four sizes of the Qwen3 series—0.6B, 1.7B, 4B, and 8B—running at FP16 precision, greedy decoding, a single frozen prompt, and no fine-tuning.74% similarUnverifiedIn a model comparison test, qwen2.5:3b correctly understood pace, produced 0 hallucinated place names, had a length error of 0.3%, peak memory of 259 MB, and took 106 seconds.73% similarUnverifiedThe dataset covers four open-weight models — Qwen2.5-7B, Mistral-7B-v0.3, Phi-3.5-mini, and Qwen2.5-1.5B — with 30 trajectories each.70% similarUnverifiedQwen3 is available on Ollama in multiple sizes including 0.6B, 1.7B, 4B, and 8B69% similarUnverifiedQwen2.5-3B-Instruct, Phi-3.5-mini (3.8B), and Llama-3.2-3B-Instruct are the three models with the best overall performance at small parameter scales at the end of 2024.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/999882API
curl https://kongchang.com/api/v1/knowledge/claims/999882MCP
get_claim(id=999882)