Holotron4-30B-A3B
Holotron4-30B-A3B is available now for rag, agents, and long context with open-weight and 256K context; evaluate it while provider pricing coverage matures.
- Family
- Holotron
- Released
- 2026-09-28
- Context
- 256K
- Parameters
- 30B (3B active)
- Architecture
- MoE + SSM Hybrid
- Specialization
- computer-use
- Openness
- Open weights
- License
- NVIDIA Open Model AgreementPermitted
- Weights
- Available
- Code
- Unknown
- Training
- Reinforcement Learning
No tracked provider pricing is available yet.
No tracked provider pricing is available yet.
About
Holotron4-30B-A3B, which H Company calls Holotron4 Nano, is a 30B-parameter (3B active) computer-use vision-language model released September 28, 2026 with the Holo4 series as the follow-up to Holotron-3-Nano. H Company applied its Holo4 post-training recipe to NVIDIA Nemotron 3 Nano Omni (NemotronH Nano Omni architecture), so the model acts through on-screen clicks and typing, code, and MCP or API tool calls. It has a 262,144-token context. BF16 and FP8 weights are on Hugging Face under the NVIDIA Open Model Agreement; it is not served on the H Models API.
Providers
No active provider routes are tracked for this model yet.
Capabilities
Benchmark peer barsfor RAG
No task-mapped benchmark peers are available for this model yet.
Migration checks
No linked migration route is available for this model yet.
API versions
Hcompany/Holotron4-30B-A3BHcompany/Holotron4-30B-A3B-FP8Holotron4 Nano