Qwen3.8-Flash-Next · IQ3_XXS
This page aggregates 1 real-world runs of Qwen3.8-Flash-Next (IQ3_XXS) on NVIDIA RTX 4090 with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.
15.9 tok/s
Decode
Decode speed
119.3 tok/s
Prefill
Prefill speed
0.84 s
TTFT
Time to first token
24 GB
VRAM
VRAM usage
1
Measured runs
1
Independent sources
Other
Source platforms
Today
Last verified
Performance
- Strata · Q4_K_M Decode 9.2 · Prefill 69 ·
- Strata · IQ3_XXS (current)Decode 15.9 · Prefill 119.3 ·
- Strata · IQ3_S Decode 13.8 · Prefill 103.5 ·
- Strata · IQ4_XS Decode 10.4 · Prefill 78 ·
- llama.cpp · Q4_K_M Decode 7.2 · Prefill 54 ·
- llama.cpp · IQ3_XXS Decode 12.4 · Prefill 93 ·
- llama.cpp · IQ3_S Decode 10.8 · Prefill 81 ·
- llama.cpp · IQ4_XS Decode 8.1 · Prefill 60.8 ·
- Strata · IQ2_XS Decode 16.6 · Prefill 124.5 ·
Core figures
- Decode (avg)
- 15.9 tok/s
- Prefill (avg)
- 119.3 tok/s
- TTFT (avg)
- 0.84 s
- VRAM (avg)
- 24 GB
- MTP acceptance rate
- —
- TTFB
- — GB
- Power draw
- — W
Configuration
Member-level fields are taken from the most recent run
- Model
- Qwen3.8-Flash-Next
- Quantization
- IQ3_XXS
- Framework
- Strata
- Version
- 0.5.2
- Context length
- 4096 tokens
- Batch size
- 512
- GPU layers
- 99
- Flash Attention
- On
Hardware
Nominal and measured figures are shown side by side; whether it runs is the reader's call
- GPU
- NVIDIA RTX 4090
- Nominal VRAM
- 24 GB
- Measured VRAM (avg)
- 24 GB
- OS
- Linux
- Driver
- 570.65
- CUDA
- 12.6
Get started
Original model resources
Base model resources may not match this quantization. Use the verified deployment weights above when available.
Sources & evidence
1 measured records in total, each traceable to its original source
0.5.2 · Linux · CUDA 12.6 · 4096 ctx
15.9 tok/s
Decode
119.3 tok/s
Prefill
0.84 s
TTFT
24 GB
VRAM
—
MTP
— W
Power draw
L0 理论估算（add-1000-records-batch3）