Decision models · Liquid AI
d1 3B
- Task
- Classifying and extracting
- Returns
- Probability or score
- Precision tier
- 82 / 100
- Size
- 3.1B
- Max input
- 32K tokens
- License
- lfm1.0
- Released
- 2026-10
What it decides
- Returns
- A calibrated probability for each typed question you pass (yes/no, one of named options, or a 2 to 10 level score), all in one forward pass with no generated text.
- Checks
- Routing and triage, moderation, intent and topic, extraction checks, reranking, rubric scoring and images
- Languages
- 16 languages
- Precision
- 82 / 100 · Decision Index 0.2.1 score 48.57, the best under 10B and ahead of Decider 35B-A3B (47.11); 69.3 on Fast Decisions (model card).
- Clarity
- Probability or score (15 of 15 clarity points)
- How it runs
- Classifier head · a small language model, so it runs in the same apps as LLMs.
- License
- lfm1.0 (Custom (read the terms))
How to run it
Quickest · Ollama
One command downloads a ready-to-run version and starts a chat.
ollama run d1:3bQuantized files · GGUF
Works in LM Studio, llama.cpp, Jan and Ollama. No sign-in needed.
huggingface.co/LiquidAI/d1-3B-GGUF (opens in a new tab)
ollama run hf.co/LiquidAI/d1-3B-GGUFOriginal weights · Hugging Face
Full-precision files from Liquid AI, for Transformers, Sentence Transformers, Text Embeddings Inference, Infinity or vLLM.
Apps that run d1 3B
Our pick for each kind of computer, assuming the model fits in its memory. Pick your own GPU or device in the Model Finder for an exact answer.
Mac (Apple silicon)
Windows PC, NVIDIA GPU
Windows PC, AMD Radeon
Linux, NVIDIA GPU
Linux server, 80 GB+ GPUs
| App | Type | Opens it as | Systems | Engine | Start |
|---|---|---|---|---|---|
| Sentence Transformers | Library | Original weights | macOSWindowsLinux |
|
Memory and speed
Each quantization on an NVIDIA GeForce RTX 4090, for one decision on a 256-token input. The best hardware table below covers other systems and devices.
| Format | Memory needed | Time per decision | Decisions per second | Quality | Download |
|---|---|---|---|---|---|
Q4_K_M | 2.5 GB | 20 ms | 50/s | 77Great quality (A) · 77 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 90/100 (quality, fit, speed and memory use) sets the ranking. | GGUF (opens in a new tab) |
Q5_K_M | 2.8 GB | 20 ms | 49/s | 79Great quality (A) · 79 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 90/100 (quality, fit, speed and memory use) sets the ranking. | GGUF files (opens in a new tab) |
Q6_K | 3.2 GB | 21 ms | 47/s | 81Great quality (A) · 81 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. | GGUF files (opens in a new tab) |
Q8_0 | 3.9 GB | 22 ms | 45/s | 82Great quality (A) · 82 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. | GGUF (opens in a new tab) |
F16 | 6.9 GB | 28 ms | 35/s | 82Great quality (A) · 82 of 100: how capable the model is at F16, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. | GGUF (opens in a new tab) Original weights (opens in a new tab) |
Best hardware for d1 3B
The least expensive GPU in our catalog that runs it without offloading is the Intel Arc A380 (launch price $139) at Q8_0.
| Quality | Hardware | Quant | Memory | Speed | Runs |
|---|---|---|---|---|---|
| 82Great quality (A) · 82 of 100: how capable the model is at F16, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. |  AMD Radeon RX 7900 XTX |
F16 | 8.1 GB / 24.0 GB | 28 ms | Runs great |
| 82Great quality (A) · 82 of 100: how capable the model is at F16, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. |  AMD Radeon RX 9070 XT |
F16 | 7.7 GB / 16.0 GB | 37 ms | Runs great |
| 82Great quality (A) · 82 of 100: how capable the model is at F16, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. |  AMD Ryzen AI Max+ 395 (128 GB) |
F16 | 14.6 GB / 96.0 GB | 75 ms | Runs great |
| 82Great quality (A) · 82 of 100: how capable the model is at F16, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. |  Apple M2 Ultra (192 GB) |
F16 | 18.4 GB / 144 GB | 67 ms | Runs great |
| 82Great quality (A) · 82 of 100: how capable the model is at F16, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. |  Apple M3 Ultra (256 GB) |
F16 | 22.2 GB / 192 GB | 65 ms | Runs great |
| 82Great quality (A) · 82 of 100: how capable the model is at F16, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. | Apple M3 Ultra (512 GB) |
F16 | 37.6 GB / 384 GB | 65 ms | Runs great |
| 82Great quality (A) · 82 of 100: how capable the model is at F16, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. |  Apple M4 Max (128 GB) |
F16 | 14.6 GB / 96.0 GB | 97 ms | Runs great |
| 82Great quality (A) · 82 of 100: how capable the model is at F16, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. | Apple M4 Max (64 GB) |
F16 | 10.7 GB / 48.0 GB | 97 ms | Runs great |
| 82Great quality (A) · 82 of 100: how capable the model is at F16, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. |  Intel Arc B580 |
F16 | 7.5 GB / 12.0 GB | 88 ms | Runs great |
| 82Great quality (A) · 82 of 100: how capable the model is at F16, whatever it runs on. Match score 91/100 (quality, fit, speed and memory use) sets the ranking. |  NVIDIA GB10 (128 GB) |
F16 | 16.1 GB / 115 GB | 79 ms | Runs great |
Estimates come from published hardware specs and our open formulas, not from running the model on your machine. Real speed varies with runtime, drivers and settings. Numbers marked “Measured” come from a linked benchmark.