GrootTeam/Groot-238B-A18B-261023
just landed · new, unvettedmodel
New, unvetted: created or first seen in the last few days, not from a trusted org and not past the signal threshold yet.
Groot is a family of open-weight reasoning models for complex real-world tasks. Mixture-of-experts, 243B total, 18B active (216 experts, 8 per token). Context: 1M tokens (config). License: mit. Vendor-reported: "It is smaller than GLM-5.3-Flash (320B) and DeepSeek-V4-Flash-0731 (284B), and it beats both on six agentic benchmarks."
- Vendor-reported: It is smaller than GLM-5.3-Flash (320B) and DeepSeek-V4-Flash-0731 (284B), and it beats both on six agentic benchmarks.
- Org
- GrootTeam
- Released
- 2026-10-23
- Status
- just landed · new, unvetted
- License
- mit (permissive)
- Parameters
- 243.4B total, 18.0B active
- Architecture
- MoE · 216 experts, 8/token · Glm5NextForConditionalGeneration
- Context
- 1,048,576 tokens
- Task
- image-text-to-text · vision
- Signal
- 2 HF likes · 46 downloads
Runs on
mixed: real files + estimate; need = weights + FP16 KV cache for 8K tokens + 1GB. A sizing calculation, not tested.
- Q4 ~153.6 GB: 2× DGX Spark (256GB), Mac 256GB, CPU only, 256GB RAM (slow)
- Q8 ~260.0 GB (estimated size): Mac 512GB
- FP16 ~488.2 GB (estimated size): no listed device
Links
First published here 2026-10-23 · updated 2026-10-23 · markdown · JSON