Eval request: new architectures (Qwen3-Coder-Next, Step-3.5-Flash, LongCat-Flash-Lite...)
#545
by
Pentium95 - opened
Base non-reasoning:
- https://huggingface.co/jdopensource/JoyAI-LLM-Flash
- https://huggingface.co/allenai/Olmo-3.1-32B-Instruct
Base Thinking / Reasoning:
- https://huggingface.co/stepfun-ai/Step-3.5-Flash (also non-reasoning)
- https://huggingface.co/upstage/Solar-Open-100B
- https://huggingface.co/arcee-ai/Trinity-Mini
- https://huggingface.co/miromind-ai/MiroThinker-1.7-mini
- https://huggingface.co/LGAI-EXAONE/K-EXAONE-236B-A23B
- ̶h̶t̶t̶p̶s̶:̶/̶/̶h̶u̶g̶g̶i̶n̶g̶f̶a̶c̶e̶.̶c̶o̶/̶Q̶w̶e̶n̶/̶Q̶w̶e̶n̶3̶.̶5̶-̶3̶9̶7̶B̶-̶A̶1̶7̶B̶
- ̶h̶t̶t̶p̶s̶:̶/̶/̶h̶u̶g̶g̶i̶n̶g̶f̶a̶c̶e̶.̶c̶o̶/̶S̶e̶r̶v̶i̶c̶e̶N̶o̶w̶-̶A̶I̶/̶A̶p̶r̶i̶e̶l̶-̶1̶.̶6̶-̶1̶5̶b̶-̶T̶h̶i̶n̶k̶e̶r̶
Finetunes non-reasoning:
- https://huggingface.co/ConicCat/Role-mo-V2-32B (chatML. Olmo-3.1-32B-Instruct finetune)
- https://huggingface.co/BirdToast/olmo-v2-stage3-lexifreak-heretic-v2 (chatML. Olmo-3.1-32B-Instruct finetune)
- https://huggingface.co/MuXodious/Olmo-3.1-32B-Instruct-impotent-heresy
- https://huggingface.co/Shifusen/Qwen3-Next-80B-A3B-Instruct-Decensored
- https://huggingface.co/rpDungeon/Qwen3-VL-32B-Heretic-v2
Finetunes Thinking / Reasoning:
- https://huggingface.co/Kilinskiy/Step-3.5-Flash-Ablitirated (also non-reasoning) (very promising)
- https://huggingface.co/hell0ks/Solar-Open-100B-jailbreak
- https://huggingface.co/Ex0bit/Step-3.5-Flash-PRISM-PRO (private, idk if possible to eval, there is a GGUF version here: https://huggingface.co/Ex0bit/Step-3.5-Flash-PRISM )
- https://huggingface.co/cerebras/Step-3.5-Flash-REAP-121B-A11B (also non-reasoning)
- https://huggingface.co/cerebras/Step-3.5-Flash-REAP-149B-A11B (also non-reasoning)
I'd love to see the Step-3.5-Flash-PRISM-PRO one for sure.