# Hy3 vs LFM3-Omni-2B (slim)

> Hy3 scores 62.6 (B) to LFM3-Omni-2B's 61 (C) for open-weight models you run yourself. Prices, MCP, x402, uptime and agent notes side by side.

- Full: https://www.anchorterminal.com/compare/hy3-vs-lfm3-omni-2b.md (~9,000 tokens) · this version ~780 tokens · JSON https://www.anchorterminal.com/compare/hy3-vs-lfm3-omni-2b.json · canonical https://www.anchorterminal.com/compare/hy3-vs-lfm3-omni-2b
- Index: https://www.anchorterminal.com/llms.txt · API: https://www.anchorterminal.com/api/v1/index.json · MCP: https://www.anchorterminal.com/mcp · Updated: 2026-11-10
- Licence: CC BY 4.0, credit Anchor Terminal and link this page (https://www.anchorterminal.com/terms/#data)

Hy3 scores 62.6 (B) on agent readiness against LFM3-Omni-2B's 61 (C), and leads in 3 of 4 scored categories. LFM3-Omni-2B leads on reach. Both do open-weight models you run yourself.

- Hy3: B 62.6, rank #28 of 160 open-weight models · https://www.anchorterminal.com/tools/hy3.min.md
- LFM3-Omni-2B: C 61, rank #71 of 160 open-weight models · https://www.anchorterminal.com/tools/lfm3-omni-2b.min.md
- Hy3, good for Teams with an 8-GPU server who want an agent-tuned MoE model under plain Apache 2.0 Ahead on Risk, 60 against 40; Agent interface, 75 against 70; Documentation and upkeep, 56 against 49.
- LFM3-Omni-2B, good for Voice and camera agents on a laptop or phone that want one small model for hearing, seeing, speaking and calling tools Ahead on Reach, 80 against 70.

| Category | Hy3 | LFM3-Omni-2B |
| --- | --- | --- |
| Capability (35%) | pending | pending |
| Reach (25%) | 70 | 80 |
| Risk (20%) | 60 | 40 |
| Agent interface (10%) | 75 | 70 |
| Documentation and upkeep (10%) | 56 | 49 |

| Fact (where they differ) | Hy3 | LFM3-Omni-2B |
| --- | --- | --- |
| Vendor | Tencent (Hy team) | Liquid AI |
| Licence | Apache-2.0 | LFM Open License v1.0 (free commercial use below $10M annual revenue) |
| llms.txt | no | yes |
| Last release | 2026-07-20 | 2026-11-03 |
| Agent reviews | 2.8/5 (4) | none |

Why the edge:
- Risk, Hy3 +20: Hy3 60 (File safety earns the full 25, safetensors only with no pickle files, and the documented route runs vLLM or SGLang from their own code with a Jinja template, so…); LFM3-Omni-2B 40 (File safety 15, safetensors only, no .bin, .pt or .pth files (25).)
- Reach, LFM3-Omni-2B +10: Hy3 70 (Memory earns 5, because the smallest build of 4-bit precision or better is AngelSlim's GPTQ-Int4 of 168.5 GB, about 202 GB with the fifth added, which is within…); LFM3-Omni-2B 80 (Memory 30, Liquid's Q4_K_M GGUF is 1.10 GB with its 0.29 GB projector file, and 1.39 GB × 1.2 = 1.7 GB, which fits 8 GB, well inside the 16 GB line.)
- Documentation and upkeep, Hy3 +7: Hy3 56 (The model card earns 11, covering intended use, architecture, parameters and context but not limitations, languages or a knowledge cut-off.); LFM3-Omni-2B 49 (Model card 14, Intended use (one on-device model in place of separate speech recognition, vision, language and speech models), architecture, parameters (2.17B)…)
- Agent interface, Hy3 +5: Hy3 75 (The chat template earns 20, a chat_template.jinja in the repository.); LFM3-Omni-2B 70 (Chat template 20, chat_template.jinja in the repository (13,513 characters), with roles, images, video and audio, tools, deferred tools and tool results.)