This is a quantized MLX pack of ukisai/Swift1.5-Qwen3.8-Flash-Next — UkisAI's reasoning-efficient finetune of Qwen3.8-Flash-Next — produced by applying the MTPLX "Optimized Speed" quantization recipe from Youssofal/Qwen3.8-Flash-Next-MTPLX-Optimized-Speed to the Swift-1.5 weights. It was forged locally with mtplx forge (MTPLX 2.12.2, qwen4-exp lane) from source revision 0bd4fe22431372cdad1979267d3ab45aa7e6150a, full-load verified before stamping, and runs with MTP draft-acceptance active and the n-gram table offloaded to SSD — the same target class as the original pack: M5 Ultra 96 GB or M5 Max 128 GB Macs with fast NVMe storage. Note vs the original Youssofal pack: this build runs qsa8bit…
agri-31b-v1-NVFP4 is an open-weight model for image and text to text from Toil Kim, released under Apache License 2.0. It has 20.9B parameters. At 16-bit it needs about 50.1 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index.
Parameters20.9B
Context—
Weights—
Licenseapache-2.0
AccessOpen weights
Monthly Downloads—
Runs On
What it takes to serve agri-31b-v1-NVFP4 (20.9B parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.
| Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits |
|---|---|---|---|---|---|
| 16-bit | 41.7 GB | 50.1 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 20.9 GB | 25.0 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 10.4 GB | 12.5 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Oct 23, 2026.
agri-31b-v1-NVFP4 on every accelerator the SAVRN Index prices, at every precision
Model Card
Identity and Version
- Repository
- toiru99/agri-31b-v1-NVFP4
- Publisher
- Toil Kim
- Task
- Image and text to text
- Modality
- Image and text
- Library
- transformers
- Parameters
- 20.9B parameters
- Languages
- ko
- Revision
- 79f4e0742a7cd930305d8274177eb9d81f9cc5e2
- First published
- 2026-10-23
- Last updated
- 2026-10-23
License and Download
- License
- apache-2.0
- Access
- Open weights, no gate
Download from Toil Kim
Released by Toil Kim through its official repository on Hugging Face. Read the license.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| 16-bit | 41.7 GB |
| 8-bit | 20.9 GB |
| 4-bit | 10.4 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About agri-31b-v1-NVFP4
How much GPU memory does agri-31b-v1-NVFP4 need?
About 50.1 GB at 16-bit and 12.5 GB at 4-bit: the weights (20.9B parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run agri-31b-v1-NVFP4 on?
At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.
Can I use agri-31b-v1-NVFP4 commercially?
Yes. agri-31b-v1-NVFP4 is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.
Similar Models
2.5x faster throughput than other NVFP4 quants. This is an Unsloth NVFP4 quantized checkpoint calibrated on a mixture of our Unsloth dataset + UltraChat dataset. Works on a 24GB VRAM GPU. Benchmarks on 1xB200 128 concurrency. For accuracy benchmarks, we conducted MMLU-Pro, AIME 2025, GPQA for FP8, BF16, NVIDIA's NVFP4 and our NVFP4s - we show our faster quants do similarly on all: Read all benchmarks in our NVFP4 blog To install vLLM in a separate venv: Then to serve the 27B NVFP4 quant: Also do NOT use the Marlin backend since it's 2x slower - use the native vLLM or cute-DSL / CUTLASS / flashinfertrtllm backends! You must use the below or you will get 2x slower inference! This checkpoint…
This model is an updated quantized version of Qwen/Qwen3.8-27B, using a mixed-precision FP4/FP8 scheme with an unquantized language-model head and updated quantization scales. See Evaluation for accuracy results. This model was produced by applying mixed-precision quantization to Qwen/Qwen3.8-27B. MLP projections are quantized to FP4, attention projections and the final MLP layers are quantized to FP8, and the KV cache is quantized to FP8, while the language-model head is kept in full precision to preserve output quality. The quantization scales were updated by calibrating on a 512-sample subset of the perfectblend dataset with a recipe that combines AWQ and GPTQ. Only the weights and…
EXL3 quantization (2.25 bits per weight on average) of UkisAI's Swift 1.5 Qwen3.8-Flash-Next (MoE, 125B total / 6B active, plus a 51B n-gram "PLE" table, MTP head and vision), made to serve the whole model on one 48 GB GPU with vLLM (continuous batching): 262,144 tokens per session without YaRN, 4 concurrent sessions sharing a 420k-token KV cache, vision and MTP. Validated on a real NVIDIA RTX PRO 5000 Blackwell (48 GB). All Swift 1.5 Qwen3.8-Flash-Next (125B-A6B) variants we publish, measured on 2026-10-14 under the same conditions (see Family comparison for every number): Download size: 66.7 GB (40.5 GB of EXL3 weights + 26.2 GB EXL3 n-gram table). vLLM users can skip the n-gram table…
EXL3 quantization (2.25 bits per weight on average) of Qwen3.8-Flash-Next (MoE, 125B total / 6B active, plus a 51B n-gram "PLE" table, MTP head and vision), made to serve the whole model on one 48 GB GPU with vLLM (continuous batching): 262,144 tokens per session without YaRN, 4 concurrent sessions sharing a 420k-token KV cache, vision and MTP. Same layout and memory footprint as our Swift 1.5 Flash-Next 2.25 bpw, which was validated on a real NVIDIA RTX PRO 5000 Blackwell (48 GB). All Qwen3.8-Flash-Next (125B-A6B) variants we publish, measured on 2026-10-14 under the same conditions (see Family comparison for every number): Download size: 66.7 GB (40.5 GB of EXL3 weights + 26.2 GB EXL3…
Solstice-AI/Qwopus3.8-27B-Flash-NVFP4-1M provides the official, production-grade NVIDIA NVFP4 release of Qwopus3.8-27B-Flash with a native-behaving 1,048,576-token (1M) context window. NVIDIA Blackwell-native NVFP4 Tensor Core quantization with baked 1M YaRN scaling, running without runtime flag overhead. Evaluated under the standardized benchmark harness: 1M YaRN Scaling & Quantization Suite: Solstice-AI