Yes, just — 0.2 GB spare

Ling 3.0 Tiny 7.9B-A1.3B at Q4_K_M fits your GeForce GTX 1660 Super entirely on the GPU at 8K context, at an estimated 107 tokens per second. Past 39K the KV cache pushes it over — quantise the cache to q8_0, or step down a quantisation, to go longer.

Fully on GPU 8K context Q4_K_M · 4.4 GB MIT Released 10 Aug 2026 Not in the Ollama library

An 8B MoE with 1.3B active and a latent KV cache on only 6 of 24 layers — reasoning and tool use sized for Apple Silicon and edge boxes.

Hardware requirements What hardware do I need to run Ling 3.0 Tiny 7.9B-A1.3B? →

At Q4_K_M with an 8K context: a 8 GB GPU (RTX 4060 Laptop) or a 16 GB Mac (M4). Minimum, recommended, high-quality and long-context tiers, with all 114 devices checked at every quantisation.

The VRAM budget

weights 4.4 GB
Weights 4.4 GB KV cache @ 8K 0.05 GB Runtime overhead 0.6 GB Free 0.2 GB of 5.3 GB

Quantisation ladder

QuantWeightsTotal @ 8KMax contextTok/sQualityFit
F16 14.7 GB 15.4 GB — ~7.8 Reference 10.1 GB over
Q8_0 7.8 GB 8.5 GB — ~21 −0.1% ppl 3.2 GB over
Q6_K 6.0 GB 6.7 GB — ~38 −0.4% ppl 1.4 GB over
Q5_K_M 5.2 GB 5.9 GB — ~61 −0.8% ppl 0.6 GB over
Q4_K_M 4.4 GB 5.1 GB 39K 107 −1.9% ppl Recommended
Q3_K_M 3.6 GB 4.2 GB 128K 132 −5.4% ppl Long context
Q2_K 3.1 GB 3.7 GB 128K 154 −15% ppl Long context

Quality is the published perplexity delta against f16 weights. Max context assumes an f16 KV cache; q8_0 roughly doubles it. Only 6 of its 24 blocks keep a per-token KV cache; the rest are linear-attention, Mamba or convolution blocks with a fixed-size state. This model uses multi-head latent attention, so its cache is a compressed latent rather than full K and V.

How to run it

terminal
$ llama-server \
    -hf inclusionAI/Ling-3.0-tiny:Q4_K_M \
    -c 8192 -ngl 99

The engine underneath most of the others. Every knob is exposed. More on llama.cpp.

01Download is 4.4 GB. Keep it on an SSD — a first load off a spinning disk takes minutes.
02Close anything else holding VRAM. A browser with hardware acceleration can sit on 1–2 GB.
03Only 0.2 GB is spare, so a long system prompt can still push it over. Its real ceiling here is 39K context.
See all models for this rig Compare with another model

Related models on a GeForce GTX 1660 Super

ModelTotalTok/s est.
Ling 3.0 Flash 124B-A5B
124B MoE · same family
Q4_K_M · 70.4 GB ~5.2 CPU offload
Gemma 4 E4B
8.0B · similar size
Q4_K_M · 5.2 GB 45 Tight fit
Llama 3.1 8B Instruct
8.03B · similar size
Q4_K_M · 6.1 GB ~25 CPU offload
Qwen3 8B
8.19B · similar size
Q4_K_M · 6.3 GB ~22 CPU offload
DeepSeek-R1-Distill-Qwen 7B
7.62B · similar size
Q4_K_M · 5.3 GB ~46 CPU offload
Qwen2.5-Coder 7B
7.62B · similar size
Q4_K_M · 5.3 GB ~46 CPU offload
Fara 7B
8.29B · similar size
Q4_K_M · 5.7 GB ~31 CPU offload

Every model against the GeForce GTX 1660 Super →

Ling 3.0 Tiny 7.9B-A1.3B on other hardware

DeviceTotalTok/s est.
GeForce RTX 4060 Laptop
7.0 GB usable · 256 GB/s
5.1 GB 82 Runs great
GeForce RTX 3060 12 GB
10.6 GB usable · 360 GB/s
5.1 GB 115 Runs great
M4 · 16 GB
10.7 GB usable · 120 GB/s
5.1 GB 43 Runs great
GeForce RTX 4090
22.4 GB usable · 1008 GB/s
5.1 GB 321 Runs great
Radeon RX 7900 XTX
22.4 GB usable · 960 GB/s
5.1 GB 306 Runs great
M3 Max · 64 GB
48.0 GB usable · 400 GB/s
5.1 GB 143 Runs great

All 114 devices, by tier →