Home / Models / Holo 3.1 35B A3B
PrismML

Holo 3.1 35B A3B

Holo 3.1 35B A3B is an open-weights 35B-parameter model by PrismML. It holds an Intelligence score of 60.7 on the local.ai leaderboard (2026-08 snapshot).

MoEopen weights
60.7
local.ai Intelligence
Parameters
35B
MoE · 3B active parameters
Benchmarks (local.ai snapshot)
τ²-bench
97%
GAIA
59%
GDPval
55%
HuggingFace metadata · updated 2026-06-26
repo
Hcompany/Holo-3.1-35B-A3B
license
apache-2.0
downloads
92,691
likes
84
Official HuggingFace repoOllamaCompare on the savings tool

Metadata refreshed from the HuggingFace API (snapshot 2026-08-15T15:53:42.405Z). Context length and long-form descriptions are the next pipeline stage.

Does it fit your hardware?

Detect or select your GPU first. Green = weights + context headroom, yellow = tight, grey = does not fit.

Q4_K_M 20.3GQ6_K 31.5GQ8_0 38.5GFP16 70G
Quantization≈ File sizeNote
IQ2_XXS9.5 GBExtreme compression, largest quality loss
IQ3_XXS12.6 GBVery small
Q3_K_M16.1 GBSmall
NVFP417.5 GBNVIDIA 4-bit
Q4_K_S18.9 GBCommon
Q4_K_M20.3 GBMost common sweet spot
Q5_K_M26.6 GBBalanced
Q6_K31.5 GBNear-lossless
Q8_038.5 GBNear-lossless
FP1670 GBOriginal weights

Actual files on HuggingFace

FileSize
model-00001-of-00026.safetensors3.72 GB
model-00006-of-00026.safetensors3.69 GB
model-00016-of-00026.safetensors3.68 GB
model-00008-of-00026.safetensors3.68 GB
model-00010-of-00026.safetensors3.68 GB
model-00018-of-00026.safetensors3.68 GB
GGUF: Hcompany/Holo-3.1-35B-A3B-GGUF · 247,643 downloads
FileSize
imatrix.gguf0.18 GBdownload
main.bf16.gguf64.61 GBdownload
mmproj.f16.gguf0.84 GBdownload
q4_k_m.gguf19.84 GBdownload
GGUF: mradermacher/Holo-3.1-35B-A3B-GGUF · 4,067 downloads
FileSize
Holo-3.1-35B-A3B.IQ4_XS.gguf17.64 GBdownload
Holo-3.1-35B-A3B.Q2_K.gguf12.05 GBdownload
Holo-3.1-35B-A3B.Q3_K_L.gguf16.87 GBdownload
Holo-3.1-35B-A3B.Q3_K_M.gguf15.61 GBdownload
Holo-3.1-35B-A3B.Q3_K_S.gguf14.14 GBdownload
Holo-3.1-35B-A3B.Q4_K_M.gguf19.71 GBdownload
Holo-3.1-35B-A3B.Q4_K_S.gguf18.52 GBdownload
Holo-3.1-35B-A3B.Q5_K_M.gguf23.03 GBdownload
Holo-3.1-35B-A3B.Q5_K_S.gguf22.33 GBdownload
Holo-3.1-35B-A3B.Q6_K.gguf26.56 GBdownload
Holo-3.1-35B-A3B.Q8_0.gguf34.37 GBdownload
Holo-3.1-35B-A3B.mmproj-Q8_0.gguf0.57 GBdownload

What you save self-hosting this model

Uses the model's own API price when it exists; otherwise the closest commercial equivalent.

Waiting for hardware detection…

Related models

Qwen
Qwen3.6 35B A3B
35B params · Intelligence 74.9
HCompany
Ornith 1.0 35B
35B params · Intelligence 74.1
Google
Gemma 4 31B It
31B params · Intelligence 66.2
NVIDIA
Nemotron 3 Nano 30B A3B
30B params · Intelligence 42.4

RunLocal presents third-party leaderboard data (local.ai / Exo Labs, 2026-08) with its own fit and cost estimates. Verify pricing, licenses and file sizes at the source before deploying a model.