Baichuan-M3-235B

Baichuan family · 235.0B params

1recorded run

26.1best tok/s

1verified

Evidence: 1 sourced run · 1 independent source · first recorded Sep 4, 2026. What these labels mean.

What the Baichuan-M3-235B numbers mean

The fastest verified result on our board is 26.1 output tok/s on an CMP 170HX x2 at Q3_K_M with a 32768-token context. That is about 38.3 ms per generated token — 500 tokens takes roughly 19.2 seconds. Prompt processing on the same record runs at 309.3 tok/s, so a 2,000-token prompt is read in about 6.5 seconds. Peak VRAM was measured at 118.0 GB.

How much VRAM does Baichuan-M3-235B need?

Baichuan-M3-235B has 235.0B parameters. These estimates use the same formula as the VRAM calculator: weights, plus about 15% for a moderate-context KV cache and runtime overhead, plus a small engine floor.

Quantisation	Estimated VRAM	Smallest card that fits
Q4_K_M (4-bit)	152.8 GB	192 GB
Q8_0 (8-bit)	287.9 GB	multi-GPU or CPU offload
FP16 (16-bit)	541.3 GB	multi-GPU or CPU offload

Long contexts and large batches raise the real number — read the smallest card as "possible with a reduced context". Measured throughput is on the benchmark board.

Fastest GPUs for this model (best tok/s)

CMP 170HX x226.1

Models with a similar footprint

Model	Params	Q4_K_M weights
MiniMax-M2.7	229B	128.8 GB
CYBER-FROST-3.8	180B	101.2 GB
Qwen3.8-Flash-Next	180B	101.2 GB
Qwen3.8-Flash-Next-hibrid48	180B	101.2 GB
Qwen3.8-Flash-Next-REAP-256-duo	180B	101.2 GB
DeepSeek-V4-Flash	291B	163.7 GB

Ranked by distance from Baichuan-M3-235B's parameter count. Full benchmark board.

All benchmarks Methodology

GPU	Model	Quant	Backend	Ctx	tok/s	prompt tok/s	VRAM GB	Source
CMP 170HX x2 · 64 GB	Baichuan-M3-235B	Q3_K_M	llama.cpp 0.28.0	32768	26.1 ✓	309.3	118.0	src