d1-Omni-2B-GGUF Multimodal Decision Model: 4GB+ VRAM, File List

At a Glance
| Item | Value |
|---|---|
| Repository | LiquidAI/d1-Omni-2B-GGUF |
| Family guide | d1-Omni guide (1 articles) |
| Publisher guide | Liquid AI: models and licenses |
| Published | 2026-10-28 |
| License | lfm1.0 |
| Formats | GGUF |
| Source type | Primary source (the publisher itself) |
Values determined by this site’s code when the information was collected. Dates are JST.
Overview
LiquidAI has released the GGUF format model of “d1-Omni-2B", a model that accepts text, image, and audio information as inputs to make decisions for specific tasks. This model is designed as a “Decision Model" that directly outputs calibrated answers in a single forward pass, rather than generating tokens sequentially like traditional generative models.
Specifications
- Architecture: LFM3 backbone, 24 layers (13 gated short convolutions, 5 sliding-window attention, 6 global attention)
- Parameter Count: 2.17B
- Inputs: Text or JSON format state, images, videos, audio clips, and written or spoken questions
- Outputs: Typed answers with probabilities (does not generate tokens or audio)
- Supported Languages: Japanese, English, Arabic, Chinese, French, German, Italian, Korean, Portuguese, Spanish, Vietnamese, Thai, Indonesian, Hindi, Russian, Polish
- License: LFM Open License v1.0
Performance and Quality
Benchmark results regarding the performance of the base model d1-Omni-2B are as follows.
→ Scroll horizontally to see all columns
| Benchmark | d1-Omni 2B | d1-omni 600M | d1 2B | d1 3B |
|---|---|---|---|---|
| Decision Index public suite (v0.3) | 49.0 | 18.3 | 51.1 | 49.0 |
| DecisionBench (Hanno Labs) v1 | 76.4 | 58.7 | 77.1 | 73.1 |
| Decision Bench (decisionbench.ai) v4 | 84.7 | 57.9 | 85.2 | 83.6 |
| Fast Decisions | 56.8 | – | 58.8 | 59.3 |
| JevBench v1.4.1 | 75.3 | 58.0 | 77.9 | 76.2 |
| ImajevBench v2.0-lite | 72.3 | 21.7 | 75.5 | 64.0 |
| WorkflowEvals | 77.1 | 48.8 | 81.6 | 74.4 |
| Typed Decisions | 63.1 | 57.5 | 61.3 | 65.0 |
| Our board (over a hundred public and private benchmarks) | 61.8 | – | 63.2 | 58.8 |
→ Scroll horizontally to see all columns
| Benchmark | d1-Omni 2B | d1-omni 600M | d1 2B | d1 3B |
|---|---|---|---|---|
| AI2D | 71.5 | 37.1 | 76.7 | 79.9 |
| BLINK | 63.5 | 37.7 | 66.2 | 59.2 |
| CountQA | 45.4 | 26.8 | 51.4 | 42.6 |
| CV-Bench | 84.0 | 40.8 | 86.1 | 82.1 |
| ERQA | 42.5 | 27.3 | 45.8 | 39.8 |
| HallusionBench | 67.5 | 52.6 | 72.1 | 65.3 |
| MMBench | 85.2 | 54.2 | 87.9 | 84.9 |
| MME | 83.9 | 61.0 | 85.4 | 82.1 |
| MMMU-Pro | 24.1 | 11.9 | 26.4 | 18.5 |
| MMRB2 | 52.6 | 48.3 | 54.2 | 49.5 |
| MMStar | 61.3 | 31.9 | 67.5 | 59.9 |
| MMVP | 81.3 | 54.0 | 82.7 | 77.0 |
| NaturalBench | 80.0 | 58.8 | 81.5 | 79.8 |
| POPE | 89.9 | 82.7 | 88.7 | 88.6 |
| V* | 76.4 | 35.6 | 76.4 | 73.8 |
| VisualPuzzles | 26.9 | 26.2 | 29.9 | 29.5 |
| VisualWebBench | 88.6 | 14.9 | 96.2 | 71.4 |
| VL-RewardBench | 68.8 | 48.1 | 69.3 | 65.0 |
| Average | 67.7 | 42.8 | 70.2 | 65.4 |
d1-Omni-2B records scores exceeding d1-omni-600M across all benchmarks. Furthermore, despite having 70% of the parameter count, it shows results outperforming d1-3B on “Our board" and 16 out of 18 image benchmarks. On the other hand, d1-2B, which lacks audio input, still maintains superior performance compared to d1-Omni-2B in most items.
Strengths and Use Cases
d1-Omni-2B specializes in advanced decision-making tasks combining audio, images, and text. Specifically, it is suited for judgment operations dependent on audio content, such as routing incoming phone calls and voice messages, or detecting acoustic events. It is also capable of multimodal information processing, such as identifying objects in images, estimating quantities, and extracting text, numbers, and dates from documents.
Regarding audio input, a key feature is that it can process audio clips and spoken questions through the same path as text and images. It is reported that the accuracy of spoken questions stays within 0.5 points of the written question score on unseen tasks. However, it is stated that accuracy for transcribing very long audio may drop significantly in non-Latin scripts. For judgments using only text and images without audio input, using d1-2B, which has higher accuracy and calibration performance, is recommended.
Hardware Requirements
Estimated requirements (calculated by Local Model Watch) — 2.2B parameters (taken from the base model LiquidAI/d1-Omni-2B)
| Your VRAM | Quantization | File size | Est. memory needed |
|---|---|---|---|
| 4GB (laptop iGPU / phone class) | BF16 | 3.2GB | 3.9GB |
Memory estimates add a 20% runtime overhead (KV cache, etc.) to the actual size of the distributed files. Actual usage varies with context length, batch size and inference engine. These figures are computed by this site from file sizes, not published by the model’s authors. Compare with other models in our VRAM quick reference. What the quantization names mean: glossary.
Can You Run It Locally?
The publisher distributes this model as GGUF.
License — lfm1.0 (Commercial use allowed with conditions): LiquidAI’s LFM Open License v1.0. Commercial use is allowed, except by legal entities with annual revenue of $10 million or more (Section 5). Redistribution requires the license text and notice of changes.
Compiled by this site’s code from the published formats and the license field. License summaries are not legal advice — check the publisher’s original terms before relying on them.
Distributed Files
Weight files published in LiquidAI/d1-Omni-2B-GGUF, listed by this site from the Hugging Face API. Sizes are the actual file sizes.
| File | Size |
|---|---|
d1-Omni-2B-BF16.gguf |
3.47GB |
d1-Omni-2B-F16.gguf |
3.47GB |
d1-Omni-2B-Q8_0.gguf |
1.85GB |
mmproj-d1-Omni-2B-BF16.gguf |
610MB |
mmproj-d1-Omni-2B-F16.gguf |
610MB |
mmproj-d1-Omni-2B-Q8_0.gguf |
326MB |
How to Get It
d1-Omni-2B-GGUF is distributed in the GGUF format. Using it requires llama.cpp, but support for d1-Omni-2B is currently awaiting the integration of the ggml-org/llama.cpp#31275 pull request. Therefore, at this time, it is necessary to build and use llama.cpp from that pull request.
You can start a server using llama-server with a command like the following:
llama-server -hf LiquidAI/d1-Omni-2B-GGUF:Q8_0 -b 8192 -ub 8192
By sending questions to the /v1/systemone endpoint, you can obtain answers for text, JSON, images, audio clips, and spoken questions. Note that multi-choice ("multiple": true), extraction, estimation, and bounding box outputs are not currently supported.
Quantized and Converted Variants
→ Scroll horizontally to see all columns
| Added | Publisher | Format | Repository | Smallest VRAM tier (build, est. memory) |
|---|---|---|---|---|
| 2026-11-08 | ggml-org | GGUF | ggml-org/d1-Omni-2B-GGUF | BF16 3.9GB (fits in 4GB VRAM) |
File sizes of each build:
- Available builds in ggml-org/d1-Omni-2B-GGUF: Q8_0 1.7GB / BF16 3.2GB
This section is appended automatically by Local Model Watch when a converted build of this model appears after publication. Memory figures are estimated from the size of the distributed files. What the quantization names mean: glossary.
Related Articles
- LFM3-Audio-2B-GGUF Audio-Text Multimodal Model: 4GB+ VRAM, File List
- d1-2B-GGUF Text, Image, and Video Decision-Making Model: 4GB+ VRAM
- LFM2.5-24B-A2B-DSpark-GGUF Text Generation Model: 4GB+ VRAM
- LFM3-Omni-2B-MLX-4bit Multimodal Model: 4GB+ VRAM
What to Read Next
- Find models by VRAM (This model runs from the 4GB tier) → Other models that run on a 8GB GPU
- Explore the same model family → d1-Omni family overview (1 articles, 1 converted builds)
- Formats this model is available in → GGUF format guide and models
- Learn about the publisher → Liquid AI: models, licenses and articles
- Other models for the same task → Other audio models

