Smaug-Agentic Model Card Released: Behavioral Fine-tune on Kimi K3
bindureddy · x · 2026-08-11
Abacus.AI has released the model card and open-source weights for Smaug-Agentic. The model is fine-tuned from Kimi K3, a 2.8T-parameter MoE model.
- Fine-tuning: Adapts behavior only without changing underlying architectural parameters, maintaining compatibility with existing Kimi K3 inference stacks.
- Behavioral Optimization: Targeted at long-horizon tool use and coding agents, such as sustaining one decisive shell action per turn over very long loops.
- License: Available under the Kimi-K3 license for deployment in coding agent frameworks.
Related event: Abacus AI Releases Smaug-Agentic Model(2 posts)→
More from coding & agent
- Malicious VS Code Extensions Deliver XWorm via Marketplace — cyb3rops · 2026-08-11
- Open-Source AI Penetration Testing Agent EVA Generates Attack Strategies and Executes Commands — tom_doerr · 2026-08-11
- After 6 Months of AI Doc Agents: Context-Gathering Beats Writing — Stock-Parking-411 · 2026-08-11
- Why I Don't Use MCP for AI in My SaaS: Architectural Tradeoffs — Hintway · 2026-08-11
- Grok Build TUI Can Integrate Third-Party Models Like DeepSeek — teortaxesTex · 2026-08-11
- Densely: Open-Source MCP for Lossless Context Compression — No_Advertising2536 · 2026-08-11