// Baichuan · Text LLM
Baichuan-M3-235B: VRAM & local deployment
235.1B (MoE) parameters. About 159 GB at 4-bit and 519 GB at BF16 with one 8K-token request.
released 2026-01-13licence: Commercial OKmemory figures are estimates
01Specs
| Publisher | Baichuan · baichuan-inc |
|---|---|
| Family / series | Baichuan · Baichuan-M3 |
| Type | Text LLM (text-generation) |
| Parameters | 235.1B (MoE) Source: counted from official safetensors weights |
| Context length | 40K tokens (40,960)max_position_embeddings in config.json |
| Architecture | 94 layers · GQA · MoE |
| Released weights | BF16 · 470 GBdownload |
| Licence | Apache-2.0 · Commercial OKPermissive licence (Apache-2.0 / MIT / CC-BY / OpenMDW): commercial use allowed. |
| Released | 2026-01-13 (Hugging Face repo created) |
02How much VRAM it needs (estimate)
Total = weights + 10% + 0.5GB overhead + KV cache for one request (FP16 KV). More concurrent users multiply the KV column.
| Precision | Weights | Total @ 8K | Total @ 32K | Smallest GPU | Mac |
|---|---|---|---|---|---|
| BF16 / FP16 (original) | 470 GB | 519 GB | 524 GB | 8 × H100 (640GB) | — |
| FP8 / INT8 / Q8_0 | 250 GB | 277 GB | 282 GB | 4 × RTX PRO 6000 (384GB) | Mac 512GB |
| 4-bit (Q4_K_M / AWQ / GPTQ-Int4) | 143 GB | 159 GB | 164 GB | 4 × RTX PRO 6000 (384GB) | Mac 256GB |
KV cache per request: 8K 1.6 GB · 32K 6.3 GB
Bits per weight: BF16 16, FP8 8.5, 4-bit 4.85. “Fits” = total ≤ 90% of memory; Mac GPU can use ~2/3 of unified memory (≤36GB) or ~3/4 (larger). For MoE, all experts must be in memory, so size by total parameters. Method →
03How to run it
Only runtimes the publisher or the runtime’s official library names for this model.
- vLLM — mentioned in the official model card
- SGLang — mentioned in the official model card
- Transformers — mentioned in the official model card
- llama.cpp — official GGUF weights
04Official repositories & quantised builds
4 official repo(s) from baichuan-inc.
- Baichuan-M3-235B main
- Baichuan-M3-235B-GPTQ-INT4 GPTQ/INT4
- Baichuan-M3-235B-FP8 FP8
- Baichuan-M3-235B-Q4_K_M-GGUF GGUF
05Other Baichuan models
All 2 →| Model | Params | 4-bit total | Released |
|---|---|---|---|
| Baichuan-M2-32B | 32.8B | 24.5 GB | 2025-08-10 |
06Sources
Data checked 2026-02-09. Parameters, context and licence are copied from the official repository; memory figures are our estimates.