huihui-ai · text

Dolphin3.0-Llama3.1-8B-abliterated

huihui-ai/Dolphin3.0-Llama3.1-8B-abliterated

Dolphin3.0-Llama3.1-8B-abliterated at Q4_K_M is exactly 4,920,745,984 bytes (4.58 GiB / 4.92 GB) — an effective 4.902 bits per weight, not the nominal 4. Its KV cache at 32K is 4.00 GiB.

From the file· summed from 1 file(s)From the file· KV per layer
Parameters
8.0B
Architecture
llama
32 layers
Context
131,072
native (config.json)
License
llama3.1

Shipped quantizations

● exact bytes, summed from published files
QuantSize●Exact bytes●Effective bpw●Tensors●Publisher
Q2_K2.96 GiB3,179,141,2483.167—RavichandranJ
Q2_K2.96 GiB3,179,142,0483.167—mradermacher
Q3_K_S3.41 GiB3,664,510,6883.651—mradermacher
Q3_K_M3.74 GiB4,018,928,5764.004—RavichandranJ
Q3_K_M3.74 GiB4,018,929,3764.004292mradermacher
Q3_K_L4.03 GiB4,321,967,8404.306—mradermacher
IQ4_XS4.18 GiB4,484,375,0724.468292mradermacher
Q4_04.34 GiB4,661,223,4244.644292RavichandranJ
Q4_K_S4.37 GiB4,692,680,7044.675—RavichandranJ
Q4_K_S4.37 GiB4,692,681,5044.675—mradermacher
Q4_K_M4.58 GiB4,920,745,9844.902—RavichandranJ
Q4_K_M4.58 GiB4,920,746,3044.902—G4E-Unknown
Q4_K_M4.58 GiB4,920,746,7844.902292mradermacher
Q5_05.21 GiB5,599,306,7525.578—RavichandranJ
Q5_K_S5.21 GiB5,599,306,7525.578—RavichandranJ
Q5_K_S5.21 GiB5,599,307,5525.578—mradermacher
Q5_K_M5.34 GiB5,733,000,1925.711292RavichandranJ
Q5_K_M5.34 GiB5,733,000,9925.711—mradermacher
Q6_K6.14 GiB6,596,020,2886.571292RavichandranJ
Q6_K6.14 GiB6,596,021,0886.571—mradermacher
Q8_07.95 GiB8,540,788,6728.509292RavichandranJ
Q8_07.95 GiB8,540,789,4728.509—mradermacher
F1614.97 GiB16,068,924,35216.008—RavichandranJ
F1614.97 GiB16,068,925,15216.008292mradermacher

KV cache by context

computed per layer
ContextKV cache (f16)●Flat formulaOverstated byFull / windowed / recurrent
4,0960.50 GiB0.50 GiB—32 / 0 / 0
8,1921.00 GiB1.00 GiB—32 / 0 / 0
16,3842.00 GiB2.00 GiB—32 / 0 / 0
32,7684.00 GiB4.00 GiB—32 / 0 / 0
65,5368.00 GiB8.00 GiB—32 / 0 / 0
131,07216.00 GiB16.00 GiB—32 / 0 / 0

Compare with

same modality, comparable size
vs gemma-4-12B-it-qat-q4_0-unquantizedvs gemma-4-E4B-itvs Qwen3-30B-A3B-Thinking-2507vs Qwen3-4Bvs Qwen3-8Bvs Llama-3.2-1B-InstructAll alternatives →Will it fit on my card? →

Will it run on your card?

full quant x context sweep
By VRAM:8GB12GB16GB24GB32GB48GB
Radeon RX 6500 XT 4GBGeForce RTX 3050 6GBGeForce RTX 5050 8GBGeForce RTX 3080 10GBGeForce RTX 2080 Ti 11GBGeForce RTX 5070 12GBGeForce RTX 5060 Ti 16GBApple M3 Pro 18GBGeForce RTX 3080 Ti 20GBGeForce RTX 5090 D V2 24GBGeForce RTX 5090 D 32GBApple M5 Max 36GBApple M5 Pro 48GBApple M5 Max 64GBApple M3 Ultra 96GBApple M5 Max 128GBApple M2 Ultra 192GBApple M3 Ultra 256GBApple M3 Ultra 512GB

Why other calculators give a different number

A parameters × bits ÷ 8 estimate puts Q4_K_M at roughly 4.21 GiB. The real file is 4.58 GiB, because a quantization is a mixture and some tensors are always kept at higher precision.

Architecture

from config.json
Layers
32
Attention heads
32
KV heads
8
Head dim
128
Hidden size
4096
Vocab
128,258
Sliding window
none
SWA period
—
MLA
no
Experts
—
Experts per token
—
use_sliding_window
—

Questions people ask

How much VRAM does Dolphin3.0-Llama3.1-8B-abliterated need?
Q4_K_M is exactly 4,920,745,984 bytes (4.58 GiB) in weights. Add the KV cache, which depends on your context length, plus roughly half a gigabyte of runtime overhead.
How large is Dolphin3.0-Llama3.1-8B-abliterated's KV cache?
4.00 GiB at 32K context with an f16 cache, computed per layer. Quantizing the cache to q8_0 roughly halves it, which is often the difference between a context length fitting and not.
Which quantization of Dolphin3.0-Llama3.1-8B-abliterated should I use?
Q4_K_M is the usual default. Pick the largest quantization that fits your card at the context you actually need — the table above gives exact sizes for every one published.