llmrun
agentica-orgQwen3ForCausalLM

DeepSWE Preview — Hardware Requirements & GPU Compatibility

Chat

DeepSWE Preview is a 32.8B-parameter open language model from agentica-org. It supports a context window of up to 40,960 tokens. At Q4_K_M it needs about 20.29 GB of VRAM — see which GPUs and Macs can run it below.

32.8B params · dense41K contextMIT1.3K downloads197 likes
Based on Qwen3 32B
B

Decent on a 24 GB GPU

Runs on24 GB Q4_K_M · ~32 tok/s · 16KMacs from24 GB Q4_K_M · ~11 tok/s

MIT

Reference GPU · RTX 4090 24 GB

Estimates, ±20% · how we estimate

Run DeepSWE Preview

One recommended way. Every other method, OS and quant is in the step-by-step guide.

Recommended

vLLM
one command · needs vLLM installed
$vllm serve agentica-org/DeepSWE-Preview --max-model-len 16384

vLLM repo with its own quantisation

Download: Weights on Hugging Face

install, OS, other quants

Other ways

each opens step-by-step

Run in cloud

Fits on RTX PRO 4500 (32 GB) (9 GB headroom) · Q4_K_M

Generation speed
~29 tok/s
generation speed
Cost per 1M output tokens
$3.29
per 1M output tokens
Deployon RunPodBest optionfrom $0.34/hr (opens in new tab)
Compare GPUs →
or

How Much VRAM Does DeepSWE Preview Need?

Select a quantization level; the listed hardware adheres to it.

QuantizationVRAM+32KOn RTX 4090 (reference)
14.6 GB19.6 GB2-bit · K-quant — 2-bit quantization with K-quant improvementsFits: 40K context · ~45 tok/s
16.6 GB21.6 GB3-bit medium — 3-bit medium quantizationFits: 40K context · ~39 tok/s
20.3 GB25.3 GB4-bit medium — 4-bit medium quantization — most popular sweet spotFits: 16K context · ~32 tok/s
24.0 GB29 GB5-bit medium — 5-bit medium quantization — good quality/size tradeoffFits: 2K context · ~27 tok/s
27.7 GB32.7 GB6-bit, very good — 6-bit quantization, very good qualityDoes not fit: needs 30 GB
33.4 GB38.4 GB8-bit, near-lossless — 8-bit quantization, near-losslessDoes not fit: needs 35.8 GB
66.2 GB71.2 GB16-bit · bfloat — Brain floating point 16 — preferred for trainingDoes not fit: needs 68.5 GB
est.calculated, no published file yet

Selected Q4_K_M

4-bit medium quantization

22.6GBat 16K context

Weights
19.7 GB
KV cache
2.7 GB
Overhead
0.3 GB

Fits RTX 4090 (reference) · 1.4 GB free · ~32 tok/s

≥50 tok/s needs ≥1.56 TB/s · RTX 5090 ~57

Which Hardware Can Run DeepSWE Preview?

Q4_K_M · 20.3 GB

DeepSWE Preview (Q4_K_M) requires 20.3 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 27+ GB is recommended. Using the full 41K context window can add up to 6.4 GB, bringing total usage to 26.7 GB. 7 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti. 122 devices with unified memory can run DeepSWE Preview, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), Mac Studio M1 Max (32 GB).

By VRAM tier · typical GPUs and Macs at Q4_K_M
8–16 GB · RTX 4060, RTX 3060 12GB, RTX 5060 Ti 16GBDoesn't fit — smallest quant needs 16.9 GB8 GB tier →12 GB tier →16 GB tier →
24 GB · RTX 4090referenceQ4_K_M · 22.6 GB · ~32 tok/s · 16K24 GB tier →
Mac 32 GB · Mac Mini M4Q4_K_M · 22.6 GB · ~4 tok/s · 40KMacs →
Mac 64 GB · Mac Studio M4 MaxQ4_K_M · 22.6 GB · ~19 tok/s · 40KMacs →

How good is DeepSWE Preview?

llmrun Score blends public benchmark boards into one number; 100 = the average open model.

Not enough data yet for an llmrun Score.

A Score needs results on enough independent boards. How the Composite Score is built

Benchmarks

This is the single board showing results for this model; badges show who ran the test.

DeepSWE Preview benchmark results

Alternatives to DeepSWE Preview

Similar models

6
ModelScoreVRAMtok/sWhy it is similar
MiniCPM V 4.7 35B A3BOpenBMB—22.1 GB~169same size
NVIDIA Nemotron 3.5 Lightning 30B A3B BF16NVIDIA—19.3 GB~197same size
Agents A1InternScience—21.5 GB~171same size
Ornith 1.0 35BDeep Reinforce—21.5 GB~171same size
EXAONE 4.5 33BLGAI-EXAONE—23.9 GB~33same size
Gemma 4 31BGoogle—22.6 GB~31same size

Specifications

Publisher
agentica-org
Parameters
32.8B
Architecture
Qwen3ForCausalLM
Context Length
40,960 tokens
Vocabulary Size
151,936
Release Date
2025-07-01
License
MIT

Resources

Frequently Asked Questions

How much VRAM does DeepSWE Preview need?

DeepSWE Preview requires 20.3 GB of VRAM at Q4_K_M, or 66.2 GB at BF16. Full 41K context adds up to 6.4 GB (26.7 GB total).

VRAM = Weights + KV Cache + Overhead

Weights = 32.8B × 4.8 bits ÷ 8 = 19.7 GB

KV Cache + Overhead ≈ 0.6 GB (at 2K context + ~0.3 GB framework)

Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.

KV Cache + Overhead ≈ 7 GB (at full 41K context)

VRAM usage by quantization

20.3 GB
26.7 GB

Learn more about VRAM estimation →

Can NVIDIA GeForce RTX 4090 run DeepSWE Preview?

Yes, at Q5_K_M (24.0 GB) or lower. Higher quantizations like Q6_K (27.7 GB) exceed the NVIDIA GeForce RTX 4090's 24 GB.

What's the best quantization for DeepSWE Preview?

For DeepSWE Preview, Q4_K_M (20.3 GB) offers the best balance of quality and VRAM usage. Q5_K_M (24.0 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 14.6 GB.

VRAM requirement by quantization

Q2_K
14.6 GB
Q4_K_M ★
20.3 GB
Q5_K_M
24.0 GB
Q6_K
27.7 GB
Q8_0
33.4 GB
BF16
66.2 GB

★ Recommended — best balance of quality and VRAM usage.

Learn more about quantization →

Can I run DeepSWE Preview on a Mac?

Yes — Mac Mini M5 Pro (24 GB) and 106 other Macs can run DeepSWE Preview. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 20.3 GB of usable unified memory (RAM minus macOS overhead).

Can I run DeepSWE Preview locally?

Yes — DeepSWE Preview can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.

How fast is DeepSWE Preview?

At Q4_K_M, DeepSWE Preview can reach ~237 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~32 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.

tok/s = (bandwidth GB/s ÷ model GB) × efficiency

Example: NVIDIA B200 → 8000 ÷ 20.3 × 0.65 = ~256 tok/s

Estimated speed at Q4_K_M (20.3 GB)

~256 tok/s
~32 tok/s
~256 tok/s
~237 tok/s

Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.

Learn more about tok/s estimation →

What's the download size of DeepSWE Preview?

At Q4_K_M, the download is about 19.66 GB. The full-precision BF16 version is 65.52 GB. The smallest option (Q2_K) is 13.92 GB.

Which GPUs can run DeepSWE Preview?

7 consumer GPUs can run DeepSWE Preview at Q4_K_M. Top options include NVIDIA GeForce RTX 5090, AMD Radeon RX 7900 XTX, NVIDIA GeForce RTX 3090. 1 GPU has plenty of headroom for comfortable inference.

Which devices can run DeepSWE Preview?

122 devices with unified memory can run DeepSWE Preview at Q4_K_M (20.3 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.