squ11z1/claude-oss-350m

TEXT GENERATIONConcurrency Cost:1Model Size:0.35BQuant:BF16Ctx Length:32kPublished:Apr 2, 2026License:otherArchitecture:Transformer0.0K Cold

Claude OSS 350M is a 0.35 billion parameter assistant model developed by squ11z1, designed to emulate a Claude-style conversational tone and interaction pattern. This compact model, with a 32768 token context length, is fine-tuned on open-source datasets for assistant behavior and consistent identity. It is primarily intended for edge deployment, low-memory experimentation, and lightweight assistant tasks requiring fast local inference.

Loading preview...

Claude OSS 350M: Edge-Optimized Assistant

Claude OSS 350M is an independent open model project by squ11z1, aiming to replicate a familiar Claude-style conversational tone and interaction within a compact, edge-sized model. This 0.35 billion parameter model focuses on delivering a consistent identity and assistant behavior in resource-constrained environments.

Key Capabilities

  • Claude-style Interaction: Designed to capture the habitual Claude-style tone and conversational patterns.
  • Lightweight & Efficient: A 350M-class model optimized for low-memory and edge deployment scenarios.
  • Instruction Following: Fine-tuned on approximately 200,000 rows of open-source data to emphasize assistant behavior and instruction adherence.
  • Multilingual Support: Capable of compact multilingual interaction.

Good For

  • Edge deployment and low-memory experimentation
  • Lightweight assistant tasks requiring fast local inference
  • Applications needing a compact model with a consistent conversational identity