tjarvis91/vfai-x-3.5-9b-options

VISIONPricing:Input $0.431 / Cached $0.0862 / Output $1.12Concurrent Unit Cost:1Model Size:9BQuant:FP8Context Size:32kTool Calling:SupportedPublished:May 12, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

The tjarvis91/vfai-x-3.5-9b-options model is a 9 billion parameter vision-language model derived from the Qwen3.5-9B architecture, developed by tjarvis91. This model represents historical research weights specifically for experiments involving chart and market-context analysis. It is designed for multimodal inputs, leveraging its Qwen3.5 lineage for conditional generation tasks.

Loading preview...

VFAi-X Legacy Vision-Language Lineage

The tjarvis91/vfai-x-3.5-9b-options model is a 9 billion parameter vision-language model, a historical research artifact derived from the Qwen3.5-9B foundation weights. Developed by tjarvis91, this model was specifically used for chart and market-context experiments, utilizing the Qwen3_5ForConditionalGeneration architecture.

Key Characteristics & Use

  • Architecture: Based on Qwen3.5-9B, configured for conditional generation with multimodal input capabilities.
  • Purpose: Represents legacy research weights for exploring vision-language tasks in financial or market analysis contexts.
  • Artifacts: Includes a full sharded Safetensors checkpoint, configuration, tokenizer, and processor assets.
  • Historical Context: This model is part of the VFAi-X lineage, distinct from Qovaryx's random-initialized scratch bases, and retains historical v3.7 and vfai-x-sniper-options branches as separate snapshots.
  • Research Loading: Requires a Transformers release supporting the qwen3_5 architecture and adherence to upstream Qwen documentation for multimodal input compatibility.

Important Considerations

  • Status: As of October 2026, Qovaryx (the broader project) is on the back burner, and desktop downloads are paused. These repositories serve as research references.
  • Limitations: Model output can be incorrect or unsupported by input. Labels like BUY/SELL/HOLD are experimental and not investment recommendations. Loading and inference were not re-run during the documentation update, and GPU memory/quantization requirements are not certified.
  • License: Distributed under Apache-2.0, requiring attribution to the foundation model (Qwen) and other third-party notices.

Finetunes
1 models