LFM3-Omni-2B is Liquid AI's first omni-modal model in the LFM3 family: a single end-to-end model that understands text, images, video and audio, and answers in text and speech. It is built for on-device use: one model replaces a separate ASR model, vision-language model, LLM and TTS model.
d1-Omni-2B
LiquidAI·Instruct·released 24 Oct 2026
ollama run d1-omni:2bd1-Omni-2B is a decision model for text, pictures, and audio, built on LFM3-Omni-2B. It returns calibrated answers in one forward pass, without generating tokens.
How much VRAM does d1-Omni-2B need?
At Q4_K_M d1-Omni-2B needs about 2.2 GB with an 8K context (2.2 GB at Q4, 5.1 GB at full precision). Pick your GPU to see what fits and how fast it runs.
Needs: ≈2.2 GB · Q4_K_M · 8K context — Pick a GPU to check the fit.
Advanced: context 8K · KV cache 16-bit
Smallest GPU for each quant and context (KV cache 16-bit)
| Quantisation | 4K | 8K | 16K | 32K | 64K |
|---|---|---|---|---|---|
| Q3_K_M | |||||
| Q4_K_M | |||||
| Q5_K_M | |||||
| Q6_K | |||||
| Q8_0 | |||||
| FP16 |
≤16 GB card24–32 GB48–96 GBmore
| Quantisation | Total |
|---|---|
| Q2_KHeavy lossSmallest, noticeably worse quality | ≈1.6 GB |
| Q3_K_MNoticeable lossSmall, some quality loss | ≈1.9 GB |
| Q4_K_MrecommendedSmall lossRecommended default for local use | ≈2.2 GB |
| Q5_K_MNear-losslessNear-lossless for most tasks | ≈2.4 GB |
| Q6_KNear-losslessPractically lossless | ≈2.6 GB |
| Q8_0LosslessLossless in practice, twice the size of Q4 | ≈2.7 GB |
| FP16 / BF16LosslessFull precision, as released | ≈4.3 GB |
Estimates for 8K tokens of context: weights + KV cache + runtime overhead. Speed assumes memory-bound decoding at your GPU’s bandwidth. Pick a GPU above and the fit column checks every row against it. Weights marked “file” use the real GGUF download size; “est.” rows are computed from the parameter count.
How to pick a quant
Take the biggest quant that fits with room left for context. Q4_K_M is the usual sweet spot. Below Q4 quality drops fast, above Q6 you mostly pay memory for nothing. Big models (70B+, large MoE) tolerate low quants better than small ones.
- LosslessF16, Q8_0
- Near-losslessQ6_K, Q5_K_M
- Small lossQ5_K_S, Q4_K_M, IQ4_NL, IQ4_XS, Q4_K_S
- Noticeable lossQ4_0, Q3_K_L, Q3_K_M, IQ3_M
- Heavy lossIQ3_XXS, Q3_K_S, IQ2_M, Q2_K
Typical quality loss, not measured per model. Tiers come from KL divergence against FP16 on Llama 3 8B: llama.cpp perplexity README: LLaMA 3 8B scoreboard (KLD and ΔPPL vs FP16, Wikitext-2); llama.cpp quantize tool: ΔPPL per quant type @ Llama-3-8B
How much context fits?
At 8K context, Q4_K_M needs ≈2.2 GB(1.2 GB weights + 0.4 GB KV cache + 0.6 GB overhead). Pick a GPU above to see how much context fits.
Weights + overheadKV cache
Can the RTX 4060 run d1-Omni-2B? →
Speed vs quality per quant
Smaller quants run faster and fit in less memory, but lose quality.
Fits in VRAMTight fitCPU offload✕ won't runQuality is typical KL divergence vs FP16, not measured for this model.
Real-world speed of d1-Omni-2B
Tokens per second measured by people running it at home. A GPU shows a median once 3 reports agree on the same setup. Reports far from our estimate wait for a review. Measure yours with llama-bench →
No medians yet. Be one of the first 3 to report a setup.
All reports
Run d1-Omni-2B locally
--quantQ4_K_Mdefault; pick your GPU to tailor itSoulfate24/d1-Omni-2B-Paretrix-GGUF
ollama run d1-omni:2bThe library build picks its own default quantisation.
ollama run hf.co/Soulfate24/d1-Omni-2B-Paretrix-GGUF:Q4_K_M llama-server -hf Soulfate24/d1-Omni-2B-Paretrix-GGUF:Q4_K_MDownloads the GGUF on first run. Installed through llama.app? The same flags work with `llama serve`.
Quantised versions and downloads
| Repository | Format | Quant | Size |
|---|---|---|---|
| LiquidAI/d1-Omni-2B-GGUF | GGUF | Q8_0 | 1.8 GB |
| ggml-org/d1-Omni-2B-GGUF | GGUF | Q8_0 | 1.8 GB |
| LiquidAI/d1-Omni-2B-GGUF | GGUF | BF16 | 3.5 GB |
| LiquidAI/d1-Omni-2B-GGUF | GGUF | F16 | 3.5 GB |
| ggml-org/d1-Omni-2B-GGUF | GGUF | BF16 | 3.5 GB |
| Soulfate24/d1-Omni-2B-Paretrix-GGUF | GGUF | — | — |
| danbev/d1-Omni-2B-GGUF | GGUF | — | — |
| LiquidAI/d1-Omni-2B-MLX-8bit | MLX | 8BIT | — |
| LiquidAI/d1-Omni-2B-MLX-fp16 | MLX | F16 | — |
| LiquidAI/d1-Omni-2B-ONNX | ONNX | — | — |
About d1-Omni-2B
d1-Omni-2B is a decision model for text, pictures, and audio, built on LFM3-Omni-2B. It returns calibrated answers in one forward pass, without generating tokens. • Native audio: clips and spoken questions go through the same pass as text and pictures, and spoken questions score within 0.5 points of written ones on unseen tasks. • Beyond classification: it also extracts text, numbers, and dates with their span, locates objects in pictures, and estimates quantities with an interval. • Millisecond decisions: 5.8 ms per decision on an NVIDIA H100 and 3.7 ms on an AMD MI325X, or 12.4 ms and 8.2 ms with a 10-second clip. Find more information about open d1 in our release blog post. The LFM3 family includes base, text, vision, audio, omni, and decision models. d1-Omni-2B adds audio to d1-2B.
Full model card on Hugging Face →Alternatives to d1-Omni-2B
All alternatives →LFM3-Audio-2B is a compact, fully featured audio model that runs locally on laptops and edge devices. It takes text or audio as input and responds with text, speech, or both.
Gemma is a family of open models built by Google DeepMind. Gemma 4 models are multimodal, handling text and image input (with audio supported on E2B, E4B, and 12B) and generating text output. This release includes open-weights models in both pre-trained and instruction-tuned variants.
Gemma is a family of open models built by Google DeepMind. Gemma 4 models are multimodal, handling text and image input (with audio supported on E2B, E4B, and 12B) and generating text output. This release includes open-weights models in both pre-trained and instruction-tuned variants.
Gemma is a family of open models built by Google DeepMind. Gemma 4 models are multimodal, handling text and image input (with audio supported on E2B, E4B, and 12B) and generating text output. This release includes open-weights models in both pre-trained and instruction-tuned variants.
Janus is a novel autoregressive framework that unifies multimodal understanding and generation. It addresses the limitations of previous approaches by decoupling visual encoding into separate pathways, while still utilizing a single, unified transformer architecture for processing.
d1-Omni-2B FAQ
- How much VRAM does d1-Omni-2B need?
- About 2.2 GB at Q4_K_M with an 8K context. Q8_0 needs about 3.1 GB and full precision about 5.1 GB. Longer contexts need more memory for the KV cache.
- Can d1-Omni-2B run on the RTX 4060?
- Yes. It fits at FP16 / BF16 (about 4.3 GB), at an estimated 35 tokens/s. It fits comfortably on the Apple M4 (16 GB), Radeon RX 6700 XT, RTX 3060 12GB.
- Can I use d1-Omni-2B commercially?
- Yes, with conditions. It is released under lfm1.0. Custom licence: read the model card before using it commercially.
- What is the context length of d1-Omni-2B?
- 128,000 tokens (125K). The full window needs far more memory than an 8K context.
- How do I run d1-Omni-2B with Ollama or llama.cpp?
- With Ollama: ollama run d1-omni:2b. With llama.cpp: llama-server -hf Soulfate24/d1-Omni-2B-Paretrix-GGUF:Q4_K_M, which starts an OpenAI-compatible server.
Rate this model
Data from Hugging Face, the model card and public leaderboards. Memory and speed figures are estimates; real numbers depend on the runtime, context length and drivers. Always check the licence yourself before commercial use.
0 comments