PRISM abliterated variants of Qwen 3.5 122B-A10B — multimodal MoE with hybrid attention. PRO, LITE, and GGUF formats.
Eric PRO
Ex0bit
AI & ML interests
None yet
Recent Activity
updated a Space 5 days ago
Ex0bit/tensorbend published a Space 5 days ago
Ex0bit/tensorbend updated a model 14 days ago
Ex0bit/jit-loraOrganizations
PRISM abliterated MiniMax M2.1 and M2.5 MoE — reasoning, coding, agentic. GGUF and safetensors.
-
Ex0bit/MiniMax-M2.5-PRISM-PRO
Text Generation • 229B • Updated • 1.68k • 15 -
Ex0bit/MiniMax-M2.5-PRISM-LITE
Text Generation • 229B • Updated • 1.47k • 3 -
Ex0bit/MiniMax-M2.5-PRISM-PRO-Tensors
229B • Updated • 12 • 2 -
Ex0bit/MiniMax-M2.1-PRISM
Text Generation • 229B • Updated • 1.72k • 106
PRISM abliterated StepFun Step 3.5 Flash MoE — safetensors, GGUF, NVFP4.
-
Ex0bit/Step-3.5-Flash-PRISM-PRO
Text Generation • 197B • Updated • 9 • 4 -
Ex0bit/Step-3.5-Flash-PRISM
Text Generation • 197B • Updated • 321 • 13 -
Ex0bit/Step-3.5-Flash-PRISM-PRO-GGUF
Text Generation • 197B • Updated • 107 • 5 -
Ex0bit/Step-3.5-Flash-PRISM-PRO-NVFP4
111B • Updated • 23 • 1
Allen AI OLMo 3 — abliterated 7B/32B-Think and NVFP4 quantized with 1M context.
Research papers and experimental models.
NVFP4-quantized Qwen 3 VLTO models for NVIDIA Blackwell / DGX Spark.
PRISM abliterated GLM models — GLM-4.6V vision, GLM-4.7, GLM-4.7-Flash MoE.
PRISM abliterated Moonshot Kimi K2.5 — multimodal MoE with REAP expert pruning.
PRISM abliterated NVIDIA Nemotron 3 Nano 30B-A3B — Mamba-2 hybrid MoE.
Experimental and research model variants.
PRISM abliterated variants of Qwen 3.5 122B-A10B — multimodal MoE with hybrid attention. PRO, LITE, and GGUF formats.
NVFP4-quantized Qwen 3 VLTO models for NVIDIA Blackwell / DGX Spark.
PRISM abliterated MiniMax M2.1 and M2.5 MoE — reasoning, coding, agentic. GGUF and safetensors.
-
Ex0bit/MiniMax-M2.5-PRISM-PRO
Text Generation • 229B • Updated • 1.68k • 15 -
Ex0bit/MiniMax-M2.5-PRISM-LITE
Text Generation • 229B • Updated • 1.47k • 3 -
Ex0bit/MiniMax-M2.5-PRISM-PRO-Tensors
229B • Updated • 12 • 2 -
Ex0bit/MiniMax-M2.1-PRISM
Text Generation • 229B • Updated • 1.72k • 106
PRISM abliterated GLM models — GLM-4.6V vision, GLM-4.7, GLM-4.7-Flash MoE.
PRISM abliterated StepFun Step 3.5 Flash MoE — safetensors, GGUF, NVFP4.
-
Ex0bit/Step-3.5-Flash-PRISM-PRO
Text Generation • 197B • Updated • 9 • 4 -
Ex0bit/Step-3.5-Flash-PRISM
Text Generation • 197B • Updated • 321 • 13 -
Ex0bit/Step-3.5-Flash-PRISM-PRO-GGUF
Text Generation • 197B • Updated • 107 • 5 -
Ex0bit/Step-3.5-Flash-PRISM-PRO-NVFP4
111B • Updated • 23 • 1
PRISM abliterated Moonshot Kimi K2.5 — multimodal MoE with REAP expert pruning.
Allen AI OLMo 3 — abliterated 7B/32B-Think and NVFP4 quantized with 1M context.
PRISM abliterated NVIDIA Nemotron 3 Nano 30B-A3B — Mamba-2 hybrid MoE.
Research papers and experimental models.
Experimental and research model variants.