Omni 2B
A multimodal model in the LFM3 series that supports audio, image, and video understanding, adding a visual encoder on top of the audio component.
Timeline (last 90 days)
Omni 2B reuses the audio components and adds a vision encoder, enabling it to process images and video, but it does not have the ability of Audio 2B to "think" before giving a text answer.
Omni 2B scores 70 on Daily Omni, while Qwen 2.5 Omni 7B scores 61.
On the tool-calling benchmark, Omni 2B scores 52, Audio 2B scores 58 and Gemma 4 E4B scores 72; tool calling is a clear weakness of LFM3.
In an 11-item transcription test covering phone numbers, invoice numbers, rupee amounts, dates and AI terms, the model got only 5 items right: invoice number, time, quantization, DLLM and Llama.cpp were correct.
In the smart-home scenario tool-calling test, Audio 2B got 9/10 right and Omni 2B got 10/10 right; both correctly handled negative instructions and multi-action commands, and neither wrongly called tools when it could not complete the task.
In the grocery bill image test, Omni 2B gave wrong change and total judgments, and failed to notice that the bill total was overprinted by 47 rupees.
Among three trap questions (blank license plate, rain-sound transcription, text injection), the model got only one right, exposing the hallucination risk that small models tend to "fill in" rather than refuse when input information is insufficient.
After 4-bit quantization, running speed is nearly 3 times faster with no drop in transcription quality; Audio's tool calling holds at 9/10, while Omni drops by 1 item; both models can run offline on a laptop, each occupying about 4 GB of memory.
All Facts (8)
Omni 2B reuses the audio components and adds a vision encoder, enabling it to process images and video, but it does not have the ability of Audio 2B to "think" before giving a text answer.
50%UnverifiedOmni 2B scores 70 on Daily Omni, while Qwen 2.5 Omni 7B scores 61.
50%UnverifiedOn the tool-calling benchmark, Omni 2B scores 52, Audio 2B scores 58 and Gemma 4 E4B scores 72; tool calling is a clear weakness of LFM3.
50%UnverifiedIn an 11-item transcription test covering phone numbers, invoice numbers, rupee amounts, dates and AI terms, the model got only 5 items right: invoice number, time, quantization, DLLM and Llama.cpp were correct.
50%UnverifiedIn the smart-home scenario tool-calling test, Audio 2B got 9/10 right and Omni 2B got 10/10 right; both correctly handled negative instructions and multi-action commands, and neither wrongly called tools when it could not complete the task.
50%UnverifiedIn the grocery bill image test, Omni 2B gave wrong change and total judgments, and failed to notice that the bill total was overprinted by 47 rupees.
50%UnverifiedAmong three trap questions (blank license plate, rain-sound transcription, text injection), the model got only one right, exposing the hallucination risk that small models tend to "fill in" rather than refuse when input information is insufficient.
50%UnverifiedAfter 4-bit quantization, running speed is nearly 3 times faster with no drop in transcription quality; Audio's tool calling holds at 9/10, while Omni drops by 1 item; both models can run offline on a laptop, each occupying about 4 GB of memory.
50%