Janai imports model with wrong name
Tried to import to Janai, latest version v0.8.3.
Model impots with the numbers letters name, fix it please.
gemini debug: Why This Specific Model Failed
With most popular models (like Llama 3 or Mistral), the developers use standardized scripts that automatically fill out this internal metadata perfectly. When Jan AI reads them, everything lines up.
However, if the creator of the ThinkingCap-Qwen3.6-27B model quantized the file manually or used a quick script that skipped tagging the general.name field, that internal ID was left blank or malformed.
Because Jan AI couldn't find a valid name inside the file, its safety mechanism kicked in: rather than crashing or showing a blank name, it grabbed the most unique identifier it had from your download link — the Hugging Face commit hash (Cf5205D6...) — and used that as the fallback name.
Essentially, the model's creator just forgot to fill out the digital name tag during the final export, forcing Jan AI to improvise!
Fixed now: I've corrected general.name on all affected files (Q4_K_M / Q5_K_M / Q6_K / Q8_0 / bf16 quants, plus the mmproj and MTP draft head) and re-uploaded. They now show as ThinkingCap Qwen3.6 27B MTP so you can tell the quants apart in Jan's list.
If you already downloaded it: no need to re-download — you can rename it inside Jan (model settings → edit name). A fresh pull will come with the correct name automatically.
Thank you will test