Models / Qwen3 32B

AI-generated content marks

This model embeds text watermarks in generated text and adds C2PA provenance metadata to supported files such as .png, .jpg, and .svg. Marks can be lost through editing, screenshots, or format conversion, so their absence does not prove a file is human-made.

Provider transparency docs →

coding reasoning writing
Parameters
32.8B
Context
128k
License
apache 2.0
Developer
Alibaba
Origin
🇨🇳 China
Released
Apr 2025

Scores

Coding
80
Reasoning
82
General
78

Score per dollar

975 pts per $/M input

general_score (78) divided by cheapest input price ($0.08/M). Higher is better value. See live pricing.

Related models

Save your hardware and every model page answers the real question: will it run on your machine, and how fast?

Join free - save your rig →

Run it locally

Per-quant memory needs and a static "can you run it?" reference - no rig entry required

Q4_K_M
20.0GB 20.0GB min 22.0GB rec
Balanced - the usual local sweet spot
Q5_K_M
24.5GB 24.5GB min 26.0GB rec
High quality, larger than Q4
Q6_K
28.5GB 28.5GB min 31.0GB rec
Near-lossless, large
Q8_0
34.0GB 34.0GB min 38.0GB rec
Near-lossless

The reference hardware

Schematic of the NVIDIA Jetson Orin NX 16GB reference rig - 16GB unified memory, 102 GB/s aggregate bandwidth
Schematic of the Single GTX 1080 Ti (11GB) reference rig - 11GB VRAM, 484 GB/s aggregate bandwidth
Schematic of the 4x H100 80GB (320GB) reference rig - 320GB VRAM, 13400 GB/s aggregate bandwidth
Schematic of the NVIDIA DGX Station 748GB reference rig - 748GB unified memory, 8000 GB/s aggregate bandwidth
Schematic of the 8x RTX 3090 rack (192GB) reference rig - 192GB VRAM, 7489 GB/s aggregate bandwidth
Schematic of the 4x RTX 5090 (128GB) reference rig - 128GB VRAM, 7168 GB/s aggregate bandwidth
Schematic of the AMD Instinct MI300X (192GB) reference rig - 192GB VRAM, 5324 GB/s aggregate bandwidth
Schematic of the 4x RTX 4090 (96GB) reference rig - 96GB VRAM, 4032 GB/s aggregate bandwidth
Schematic of the 2x RTX 5090 (64GB) reference rig - 64GB VRAM, 3584 GB/s aggregate bandwidth
Schematic of the 2x RTX 3090 (48GB) reference rig - 48GB VRAM, 1872 GB/s aggregate bandwidth
Schematic of the Single RTX 5090 (32GB) reference rig - 32GB VRAM, 1792 GB/s aggregate bandwidth
Schematic of the RTX PRO 6000 Blackwell (96GB) reference rig - 96GB VRAM, 1792 GB/s aggregate bandwidth
Schematic of the Mac Studio M4 Ultra 192GB reference rig - 192GB unified memory, 1092 GB/s aggregate bandwidth
Schematic of the Mac Studio M4 Ultra 512GB reference rig - 512GB unified memory, 1092 GB/s aggregate bandwidth
Schematic of the Single RTX 4090 (24GB) reference rig - 24GB VRAM, 1008 GB/s aggregate bandwidth
Schematic of the MacBook Pro M5 Max 128GB reference rig - 128GB unified memory, 614 GB/s aggregate bandwidth
Schematic of the Dual EPYC 9004 + 768GB DDR5-4800 reference rig - 768GB unified memory, 460 GB/s aggregate bandwidth
Schematic of the DGX Spark 128GB unified reference rig - 128GB unified memory, 273 GB/s aggregate bandwidth
Schematic of the Ryzen AI Max+ 395 128GB reference rig - 128GB unified memory, 256 GB/s aggregate bandwidth
Schematic of the Jetson AGX Orin 64GB reference rig - 64GB unified memory, 204 GB/s aggregate bandwidth
Schematic of the Epyc + 512GB DDR4-3200 + 2x RTX 3090 reference rig - 560GB unified memory, 204 GB/s aggregate bandwidth
Schematic of the Epyc + 512GB DDR4-2400 + 2x RTX 3090 reference rig - 560GB unified memory, 153 GB/s aggregate bandwidth

22 reference configs, drawn in-house. Scroll for more.

Can you run it? - reference rigs

Rig Q4_K_M Q5_K_M Q6_K Q8_0
NVIDIA Jetson Orin NX 16GB no -> cloud no -> cloud no -> cloud no -> cloud
Single GTX 1080 Ti (11GB) no -> cloud no -> cloud no -> cloud no -> cloud
4x H100 80GB (320GB) fast 368.5t/s fast 300.8t/s fast 258.6t/s fast 216.8t/s
NVIDIA DGX Station 748GB fast 220.0t/s fast 179.6t/s fast 154.4t/s fast 129.4t/s
8x RTX 3090 rack (192GB) fast 205.9t/s fast 168.1t/s fast 144.5t/s fast 121.2t/s
4x RTX 5090 (128GB) fast 197.1t/s fast 160.9t/s fast 138.3t/s fast 115.9t/s
AMD Instinct MI300X (192GB) fast 146.4t/s fast 119.5t/s fast 102.8t/s fast 86.1t/s
4x RTX 4090 (96GB) fast 110.9t/s fast 90.5t/s fast 77.8t/s fast 65.2t/s
2x RTX 5090 (64GB) fast 98.6t/s fast 80.5t/s fast 69.2t/s fast 58.0t/s
2x RTX 3090 (48GB) fast 51.5t/s fast 42.0t/s fast 36.1t/s fast 30.3t/s
Single RTX 5090 (32GB) fast 49.3t/s fast 40.2t/s fast 34.6t/s offload
RTX PRO 6000 Blackwell (96GB) fast 49.3t/s fast 40.2t/s fast 34.6t/s fast 29.0t/s
Mac Studio M4 Ultra 192GB fast 32.8t/s fast 26.7t/s fast 23.0t/s ok 19.3t/s
Mac Studio M4 Ultra 512GB fast 32.8t/s fast 26.7t/s fast 23.0t/s ok 19.3t/s
Single RTX 4090 (24GB) fast 27.7t/s offload offload offload
MacBook Pro M5 Max 128GB ok 18.4t/s ok 15.0t/s ok 12.9t/s ok 10.8t/s
Dual EPYC 9004 + 768GB DDR5-4800 ok 12.7t/s ok 10.3t/s ok 8.9t/s slow 7.5t/s
DGX Spark 128GB unified slow 7.5t/s slow 6.1t/s slow 5.3t/s slow 4.4t/s
Ryzen AI Max+ 395 128GB slow 7.0t/s slow 5.8t/s slow 4.9t/s slow 4.1t/s
Jetson AGX Orin 64GB slow 5.6t/s slow 4.6t/s slow 4.0t/s slow 3.3t/s
Epyc + 512GB DDR4-3200 + 2x RTX 3090 slow 5.6t/s slow 4.6t/s slow 4.0t/s slow 3.3t/s
Epyc + 512GB DDR4-2400 + 2x RTX 3090 slow 4.2t/s slow 3.5t/s slow 3.0t/s slow 2.5t/s

Fit tiers use the same will-it-run logic as the rig finder. For comfortable fits, the badge reflects decode speed: fast >=20 t/s, ok 8-20 t/s, slow <8 t/s. t/s is a bandwidth estimate, not a measured benchmark.

Download options

Q4_K_M official -5% vs fp16
20.0GB dl 20.0GB min 22.0GB rec
REC RAM vs largest quant
Q5_K_M official -3% vs fp16
24.5GB dl 24.5GB min 26.0GB rec
REC RAM vs largest quant
Q6_K official -1% vs fp16
28.5GB dl 28.5GB min 31.0GB rec
REC RAM vs largest quant
Q8_0 official -1% vs fp16
34.0GB dl 34.0GB min 38.0GB rec
REC RAM vs largest quant

Or run it in the cloud

Live per-provider pricing, throughput and uptime - refreshed about 8 hours ago via OpenRouter. Click a column to sort.

Provider Type Input $/M Output $/M Cache $/M Tok/s Latency Uptime Value
SiliconFlow
API 0.14 0.57 - - - 100.00% best uptime
DeepInfra
API 0.08 0.28 - - - 99.98% cheapest

Default order: throughput among 95%+ uptime providers, then latency; subscriptions last. Sort by any column. Subscription rows show $/mo in the Value column - per-token columns are "-". Affiliate links are marked sponsored / nofollow. Confirm current pricing on the provider's site before committing.

Detailed API pricing page + JSON endpoint →

See who runs Alibaba in production →

PRICE HISTORY

Inference cost over time

Data accumulates from the first daily sync - longer ranges populate over time. Prices come from OpenRouter snapshots, not a historical API.

Loading price history...