Each of the 9 Enneagram types trained on top of Gemma 3 4B.
AI & ML interests
Indistinguishable from thought.
Recent Activity
An ongoing research initiative into refining and scaling the Atom persona from 4B to 400B+
-
vanta-research/atom-v1-preview-4b
Text Generation • 4B • Updated • 2.07k • 8 -
vanta-research/atom-olmo3-7b
Text Generation • 7B • Updated • 222 • 5 -
vanta-research/atom-v1-preview-8b
Text Generation • 8B • Updated • 476 • 4 -
vanta-research/atom-v1-preview-12b
Text Generation • 12B • Updated • 254 • 5
Distinct AI personas, each optimized for different cognitive styles and use cases.
VANTA Research in the wild: Thanks for your support!
-
mradermacher/mox-tiny-1-i1-GGUF
Text Generation • 8B • Updated • 15.4k • 4 -
mradermacher/scout-4b-i1-GGUF
Text Generation • 4B • Updated • 4.16k • 3 -
mradermacher/atom-v1-preview-8b-i1-GGUF
8B • Updated • 6.92k • 2 -
sillykiwi/Gemma3-4B-CodeCenturion
Text Generation • 4B • Updated • 9 • 2
high quality datasets designed to strengthen interpretability and general communicative flow
-
vanta-research/poetic-imagery-small
Viewer • Updated • 520 • 14 • 5 -
vanta-research/excitement-small
Viewer • Updated • 388 • 22 • 4 -
vanta-research/human-ai-collaboration-1
Preview • Updated • 22 • 10 -
vanta-research/human-ai-collaboration-2
Viewer • Updated • 6.32k • 20 • 5
Mox-1: Opinionated, unconventional, yet creatively helpful
cutting edge language models optimized for high quality conversational and collaborative interaction.
Apollo series: VANTA's core reasoning models. Demonstrates iterative training methodology and persona stability.
-
vanta-research/apollo-astralis-8b
Text Generation • 8B • Updated • 1.12k • 11 -
vanta-research/apollo-v1-7b
Text Generation • Updated • 51 • 2 -
vanta-research/apollo-astralis-4b
Text Generation • 4B • Updated • 108 • 6 -
vanta-research/apollo-astralis-2
Text Generation • 9B • Updated • 2
Highly specialized versions of Atom for specific domains
Each of the 9 Enneagram types trained on top of Gemma 3 4B.
Mox-1: Opinionated, unconventional, yet creatively helpful
An ongoing research initiative into refining and scaling the Atom persona from 4B to 400B+
-
vanta-research/atom-v1-preview-4b
Text Generation • 4B • Updated • 2.07k • 8 -
vanta-research/atom-olmo3-7b
Text Generation • 7B • Updated • 222 • 5 -
vanta-research/atom-v1-preview-8b
Text Generation • 8B • Updated • 476 • 4 -
vanta-research/atom-v1-preview-12b
Text Generation • 12B • Updated • 254 • 5
cutting edge language models optimized for high quality conversational and collaborative interaction.
Distinct AI personas, each optimized for different cognitive styles and use cases.
Apollo series: VANTA's core reasoning models. Demonstrates iterative training methodology and persona stability.
-
vanta-research/apollo-astralis-8b
Text Generation • 8B • Updated • 1.12k • 11 -
vanta-research/apollo-v1-7b
Text Generation • Updated • 51 • 2 -
vanta-research/apollo-astralis-4b
Text Generation • 4B • Updated • 108 • 6 -
vanta-research/apollo-astralis-2
Text Generation • 9B • Updated • 2
VANTA Research in the wild: Thanks for your support!
-
mradermacher/mox-tiny-1-i1-GGUF
Text Generation • 8B • Updated • 15.4k • 4 -
mradermacher/scout-4b-i1-GGUF
Text Generation • 4B • Updated • 4.16k • 3 -
mradermacher/atom-v1-preview-8b-i1-GGUF
8B • Updated • 6.92k • 2 -
sillykiwi/Gemma3-4B-CodeCenturion
Text Generation • 4B • Updated • 9 • 2
Highly specialized versions of Atom for specific domains
high quality datasets designed to strengthen interpretability and general communicative flow
-
vanta-research/poetic-imagery-small
Viewer • Updated • 520 • 14 • 5 -
vanta-research/excitement-small
Viewer • Updated • 388 • 22 • 4 -
vanta-research/human-ai-collaboration-1
Preview • Updated • 22 • 10 -
vanta-research/human-ai-collaboration-2
Viewer • Updated • 6.32k • 20 • 5