Abacus.AI launches Smaug open-weight LLM model line
4 companies named across 1 category, one of 11 articles referencing WhatsApp. Previous coverage: Rogers Communications opens Rogers Satellite province-wide in British Columbia for wildfire pilot (Aug 2026).
Companies mentioned
Best suited for
- Seniority
- C Level / Executive Team
- Job function
- Chief Technology Officer
- Persona
- Platform / Infrastructure Architect
- Buyer role
- Decision Maker / Budget Holder
- Buyer journey
- Need to Buy
- Adoption curve
- Early Adopters
- Technology maturity
- Emerging Exploration
- Industry
- Information Technology / Software & Services / Internet Software & Services / Frontier AI / Foundation Models
Our classification, not the publisher's statement. Best suited for, not only for.
Abacus.AI introduced three open-weight large language models within its enterprise agentic AI platform and Super Assistant, as part of a Smaug model line. The update centers on how the models are packaged and deployed, including where they run and how teams keep control of their data.
Abacus.AI said Smaug was designed to support long-running self-improving agentic loops in a cost-efficient way. The company also described the family as configurable for in-cloud hosting, with models available to download from Hugging Face.
Smaug uses a fine-tuning approach that Abacus.AI said applies to open-source base models. The company stated the technique improved the performance of long-running agentic loops by 15–20% without increasing cost, and that the models could be hosted in an enterprise cloud VPC environment.
The company released Smaug Agentic, Smaug Flash, and Smaug Mini. Smaug Flash was described as optimized for personal agents that can connect to messaging apps including WhatsApp, Telegram, and Slack and maintain long-running conversations; Abacus.AI said it was fine-tuned on DeepSeek Flash. Smaug Agentic was described as a 2T model based on Kimi K3; Smaug Mini was described as a 27B model optimized for multimodal use cases and smaller reasoning workloads. Abacus.AI also published the models on LiveBench AI and made a comparison to base models available, along with a technical brief. “Open-weight models are rapidly closing the gap to frontier closed models, but still underperform in long-running agent loops. The Smaug line addresses this shortcoming while still being 10-100x cheaper than closed-source models,” said Bindu Reddy, CEO of Abacus.AI.
Press release, provided by Cision on behalf of Abacus.AI, Inc.. Read the original.