Can the RTX 4060 run Inkling-2-Mini?
Inkling-2-Mini·26.4B (3.5B active)·GeForce RTX 4060·8 GB VRAM
Yes. The RTX 4060 runs Inkling-2-Mini fully in VRAM at FP16 / BF16 (about 1.9 GB of 8.0 GB), at roughly 22 tokens/s.
- best quant
- FP16 / BF16
- needs
- 1.9 GB
- usable
- 8.0 GB
- est. speed
- ~22 tok/s
Estimates for an 8K context; offload assumes 32 GB of system RAM.Try other contexts on the Inkling-2-Mini page →
Every quantisation of Inkling-2-Mini on the RTX 4060
| Quant | Memory | Fit | Speed |
|---|---|---|---|
| Q2_K | 9.7 GB | offload | ~81 tok/s |
| Q3_K_M | 14 GB | offload | ~36 tok/s |
| Q4_K_M | 17 GB | offload | ~25 tok/s |
| Q5_K_M | 20 GB | offload | ~20 tok/s |
| Q6_K | 23 GB | offload | ~16 tok/s |
| Q8_0 | 1.6 GB | fits | ~41 tok/s |
| FP16 / BF16 pick | 1.9 GB | fits | ~22 tok/s |
Speed is a bandwidth-based estimate for generating text (272 GB/s); prompt processing and long contexts are slower.
Measured speed on the RTX 4060
No community reports for this pair yet.
Report your own speed →Run Inkling-2-Mini on the RTX 4060
Using FP16 / BF16, the best fit for RTX 4060.
ollama run inkling-2-miniThe library build picks its own default quantisation.
ollama run hf.co/unsloth/Inkling-2-Mini-GGUF:BF16 llama-server -hf unsloth/Inkling-2-Mini-GGUF:BF16Downloads the GGUF on first run. Installed through llama.app? The same flags work with `llama serve`.
lms get unsloth/Inkling-2-Mini-GGUF@BF16Then load it from the LM Studio app or with `lms load`.
from transformers import pipeline
pipe = pipeline("image-text-to-text", model="thinkingmachines/Inkling-2-Mini", dtype="auto", device_map="auto")
messages = [
{
"role": "user",
"content": [
{"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/bee.jpg"},
{"type": "text", "text": "Describe this image."},
],
}
]
out = pipe(text=messages, max_new_tokens=256)
print(out[0]["generated_text"][-1]["content"])Vision models that fit the RTX 4060 comfortably
All that fit →| Model | Params | Best quant | Memory | Speed |
|---|---|---|---|---|
| 9.7B | Q4_K_M | ≈7.0 GBFits in memory | ~26 tok/s | |
| 8.8B | Q5_K_M | ≈7.3 GBFits in memory | ~24 tok/s | |
| 9.4B | Q4_K_M | ≈7.2 GBFits in memory | ~26 tok/s | |
| 8.3B | Q6_K | ≈7.0 GBFits in memory | ~22 tok/s | |
| 10.3B | Q4_K_M | ≈6.8 GBFits in memory | ~24 tok/s | |
| 9.4B | Q4_K_M | ≈7.4 GBFits in memory | ~26 tok/s |
Other GPUs that run Inkling-2-Mini
Frequently asked questions
- Can the RTX 4060 run Inkling-2-Mini?
- Yes. The RTX 4060 runs Inkling-2-Mini fully in VRAM at FP16 / BF16 (about 1.9 GB of 8.0 GB), at roughly 22 tokens/s.
- How fast does Inkling-2-Mini run on the RTX 4060?
- An estimate from memory bandwidth gives about 22 tokens/s at FP16 / BF16.
- Which quantisation of Inkling-2-Mini should I use on the RTX 4060?
- FP16 / BF16: the largest one that fits. Expected quality: lossless compared with the full-precision weights.
Memory and speed are estimates from the model's size and the card's memory bandwidth; real numbers depend on the runtime, drivers and context length. Community figures are medians of user reports.