Qwen3 Coder Next 80B-A3B

Sparse coding model for repository work and coding agents.

Stored parameters 79.7B
Native context 256Ktokens
ArchitectureMixture of experts
Use in calculator

3B parameters active per token. All 79.7B stored parameters count toward memory.

Memory by precision and context

GiB · one request · FP16 cache · default single-device buffers. Extended contexts are marked *. Text inference only.

Weight precisionWeights8K total32K total128K total256K total
Q4 · smaller44.548.348.951.254.2
Q5 · medium52.957.457.960.263.2
Q8 · larger78.885.486.088.291.2
FP16 / BF16148.4160.5161.1163.4166.4

Hardware fit at Q4 and 8K

One request, FP16 cache. Multi-GPU and compressed or sliding caches require compatible runtime support.

View 199 hardware configurations
HardwareMemory neededAvailableEstimateOpen plan
GeForce RTX 50508 GiB · GPU48.3 GiB7.0 GiBOver capacity
GeForce RTX 50608 GiB · GPU48.3 GiB7.0 GiBOver capacity
GeForce RTX 5060 Ti 8 GB8 GiB · GPU48.3 GiB7.0 GiBOver capacity
GeForce RTX 5060 Ti 16 GB16 GiB · GPU48.3 GiB15.0 GiBOver capacity
GeForce RTX 507012 GiB · GPU48.3 GiB11.0 GiBOver capacity
GeForce RTX 5070 Ti16 GiB · GPU48.3 GiB15.0 GiBOver capacity
GeForce RTX 508016 GiB · GPU48.3 GiB15.0 GiBOver capacity
GeForce RTX 509032 GiB · GPU48.3 GiB31.0 GiBOver capacity
GeForce RTX 3060 12 GB12 GiB · GPU48.3 GiB11.0 GiBOver capacity
GeForce RTX 4060 Ti 16 GB16 GiB · GPU48.3 GiB15.0 GiBOver capacity
GeForce RTX 4070 Ti SUPER16 GiB · GPU48.3 GiB15.0 GiBOver capacity
GeForce RTX 309024 GiB · GPU48.3 GiB23.0 GiBOver capacity
GeForce RTX 3090 Ti24 GiB · GPU48.3 GiB23.0 GiBOver capacity
GeForce RTX 409024 GiB · GPU48.3 GiB23.0 GiBOver capacity
GeForce RTX 40608 GiB · GPU48.3 GiB7.0 GiBOver capacity
GeForce RTX 4060 Ti 8 GB8 GiB · GPU48.3 GiB7.0 GiBOver capacity
GeForce RTX 4070 GDDR6X12 GiB · GPU48.3 GiB11.0 GiBOver capacity
GeForce RTX 4070 SUPER12 GiB · GPU48.3 GiB11.0 GiBOver capacity
GeForce RTX 4070 Ti12 GiB · GPU48.3 GiB11.0 GiBOver capacity
GeForce RTX 408016 GiB · GPU48.3 GiB15.0 GiBOver capacity
GeForce RTX 4080 SUPER16 GiB · GPU48.3 GiB15.0 GiBOver capacity
GeForce RTX 5090 Laptop GPU 24 GB24 GiB · GPU48.3 GiB23.0 GiBOver capacity
GeForce RTX 5080 Laptop GPU 16 GB16 GiB · GPU48.3 GiB15.0 GiBOver capacity
GeForce RTX 5070 Ti Laptop GPU 12 GB12 GiB · GPU48.3 GiB11.0 GiBOver capacity
GeForce RTX 5070 Laptop GPU 12 GB12 GiB · GPU48.3 GiB11.0 GiBOver capacity
GeForce RTX 5070 Laptop GPU 8 GB8 GiB · GPU48.3 GiB7.0 GiBOver capacity
GeForce RTX 5060 Laptop GPU 8 GB8 GiB · GPU48.3 GiB7.0 GiBOver capacity
GeForce RTX 5050 Laptop GPU 8 GB8 GiB · GPU48.3 GiB7.0 GiBOver capacity
GeForce RTX 4090 Laptop GPU 16 GB16 GiB · GPU48.3 GiB15.0 GiBOver capacity
GeForce RTX 4080 Laptop GPU 12 GB12 GiB · GPU48.3 GiB11.0 GiBOver capacity
GeForce RTX 4070 Laptop GPU 8 GB8 GiB · GPU48.3 GiB7.0 GiBOver capacity
GeForce RTX 4060 Laptop GPU 8 GB8 GiB · GPU48.3 GiB7.0 GiBOver capacity
GeForce RTX 4050 Laptop GPU 6 GB6 GiB · GPU48.3 GiB5.0 GiBOver capacity
RTX PRO 2000 Blackwell16 GiB · GPU48.3 GiB15.0 GiBOver capacity
RTX PRO 4000 Blackwell SFF24 GiB · GPU48.3 GiB23.0 GiBOver capacity
RTX PRO 4000 Blackwell24 GiB · GPU48.3 GiB23.0 GiBOver capacity
RTX PRO 4500 Blackwell Workstation32 GiB · GPU48.3 GiB31.0 GiBOver capacity
RTX PRO 5000 Blackwell 48 GB48 GiB · GPU48.3 GiB47.0 GiBOver capacity
RTX PRO 5000 Blackwell 72 GB72 GiB · GPU48.3 GiB71.0 GiBRoom to spare
RTX PRO 6000 Blackwell Workstation96 GiB · GPU48.3 GiB95.0 GiBRoom to spare
RTX PRO 6000 Blackwell Max-Q96 GiB · GPU48.3 GiB95.0 GiBRoom to spare
RTX A600048 GiB · GPU48.3 GiB47.0 GiBOver capacity
RTX 6000 Ada Generation48 GiB · GPU48.3 GiB47.0 GiBOver capacity
RTX 5000 Ada Generation32 GiB · GPU48.3 GiB31.0 GiBOver capacity
A100 80 GB PCIe80 GiB · GPU48.3 GiB78.0 GiBRoom to spare
A100 80 GB SXM80 GiB · GPU48.3 GiB78.0 GiBRoom to spare
H100 80 GB SXM80 GiB · GPU48.3 GiB78.0 GiBRoom to spare
H100 NVL 94 GB PCIe94 GiB · GPU48.3 GiB92.0 GiBRoom to spare
H200 141 GB SXM141 GiB · GPU48.3 GiB139.0 GiBRoom to spare
H200 NVL 141 GB PCIe141 GiB · GPU48.3 GiB139.0 GiBRoom to spare
L40S48 GiB · GPU48.3 GiB46.0 GiBOver capacity
B200 180 GB SXM180 GiB · GPU48.3 GiB178.0 GiBRoom to spare
B300 288 GB SXM288 GiB · GPU48.3 GiB286.0 GiBRoom to spare
A4048 GiB · GPU48.3 GiB46.0 GiBOver capacity
Radeon RX 680016 GiB · GPU48.3 GiB15.0 GiBOver capacity
Radeon RX 6800 XT16 GiB · GPU48.3 GiB15.0 GiBOver capacity
Radeon RX 6950 XT16 GiB · GPU48.3 GiB15.0 GiBOver capacity
Radeon RX 7600 XT16 GiB · GPU48.3 GiB15.0 GiBOver capacity
Radeon RX 7700 XT12 GiB · GPU48.3 GiB11.0 GiBOver capacity
Radeon RX 7800 XT16 GiB · GPU48.3 GiB15.0 GiBOver capacity
Radeon RX 7900 GRE16 GiB · GPU48.3 GiB15.0 GiBOver capacity
Radeon RX 7900 XT20 GiB · GPU48.3 GiB19.0 GiBOver capacity
Radeon RX 7900 XTX24 GiB · GPU48.3 GiB23.0 GiBOver capacity
Radeon RX 9060 XT 16 GB16 GiB · GPU48.3 GiB15.0 GiBOver capacity
Radeon RX 9070 GRE12 GiB · GPU48.3 GiB11.0 GiBOver capacity
Radeon RX 907016 GiB · GPU48.3 GiB15.0 GiBOver capacity
Radeon RX 9070 XT16 GiB · GPU48.3 GiB15.0 GiBOver capacity
Radeon AI PRO R970032 GiB · GPU48.3 GiB31.0 GiBOver capacity
Radeon PRO W7800 32 GB32 GiB · GPU48.3 GiB31.0 GiBOver capacity
Radeon PRO W7800 48 GB48 GiB · GPU48.3 GiB47.0 GiBOver capacity
Radeon PRO W790048 GiB · GPU48.3 GiB47.0 GiBOver capacity
Instinct MI210 PCIe64 GiB · GPU48.3 GiB62.0 GiBRoom to spare
Instinct MI300X OAM192 GiB · GPU48.3 GiB190.0 GiBRoom to spare
Instinct MI325X OAM256 GiB · GPU48.3 GiB254.0 GiBRoom to spare
Instinct MI350X OAM288 GiB · GPU48.3 GiB286.0 GiBRoom to spare
Instinct MI355X OAM288 GiB · GPU48.3 GiB286.0 GiBRoom to spare
Instinct MI350P PCIe144 GiB · GPU48.3 GiB142.0 GiBRoom to spare
Arc Pro B5016 GiB · GPU48.3 GiB15.0 GiBOver capacity
Arc Pro B6024 GiB · GPU48.3 GiB23.0 GiBOver capacity
Arc Pro B6532 GiB · GPU48.3 GiB31.0 GiBOver capacity
Arc Pro B7032 GiB · GPU48.3 GiB31.0 GiBOver capacity
Arc A770 16 GB16 GiB · GPU48.3 GiB15.0 GiBOver capacity
Arc B57010 GiB · GPU48.3 GiB9.0 GiBOver capacity
Arc B58012 GiB · GPU48.3 GiB11.0 GiBOver capacity
Mac mini M1 · 8 GB (8-core GPU)8 GiB · Unified memory48.3 GiB6.0 GiBOver capacity
Mac mini M1 · 16 GB (8-core GPU)16 GiB · Unified memory48.3 GiB12.0 GiBOver capacity
Mac mini M2 · 8 GB (10-core GPU)8 GiB · Unified memory48.3 GiB6.0 GiBOver capacity
Mac mini M2 · 16 GB (10-core GPU)16 GiB · Unified memory48.3 GiB12.0 GiBOver capacity
Mac mini M2 · 24 GB (10-core GPU)24 GiB · Unified memory48.3 GiB18.0 GiBOver capacity
Mac mini M2 Pro · 16 GB (16-core GPU)16 GiB · Unified memory48.3 GiB12.0 GiBOver capacity
Mac mini M2 Pro · 32 GB (16-core GPU)32 GiB · Unified memory48.3 GiB24.0 GiBOver capacity
Mac mini M2 Pro · 16 GB (19-core GPU)16 GiB · Unified memory48.3 GiB12.0 GiBOver capacity
Mac mini M2 Pro · 32 GB (19-core GPU)32 GiB · Unified memory48.3 GiB24.0 GiBOver capacity
MacBook Air M3 · 8 GB (8-core GPU)8 GiB · Unified memory48.3 GiB6.0 GiBOver capacity
MacBook Air M3 · 16 GB (8-core GPU)16 GiB · Unified memory48.3 GiB12.0 GiBOver capacity
MacBook Air M3 · 24 GB (8-core GPU)24 GiB · Unified memory48.3 GiB18.0 GiBOver capacity
MacBook Air M3 · 8 GB (10-core GPU)8 GiB · Unified memory48.3 GiB6.0 GiBOver capacity
MacBook Air M3 · 16 GB (10-core GPU)16 GiB · Unified memory48.3 GiB12.0 GiBOver capacity
MacBook Air M3 · 24 GB (10-core GPU)24 GiB · Unified memory48.3 GiB18.0 GiBOver capacity
MacBook Pro M1 Pro · 16 GB (16-core GPU)16 GiB · Unified memory48.3 GiB12.0 GiBOver capacity
MacBook Pro M1 Pro · 32 GB (16-core GPU)32 GiB · Unified memory48.3 GiB24.0 GiBOver capacity
Mac Studio M1 Max · 32 GB (24-core GPU)32 GiB · Unified memory48.3 GiB24.0 GiBOver capacity
Mac Studio M1 Max · 64 GB (24-core GPU)64 GiB · Unified memory48.3 GiB48.0 GiBOver capacity
Mac Studio M1 Max · 32 GB (32-core GPU)32 GiB · Unified memory48.3 GiB24.0 GiBOver capacity
Mac Studio M1 Max · 64 GB (32-core GPU)64 GiB · Unified memory48.3 GiB48.0 GiBOver capacity
Mac Studio M1 Ultra · 64 GB (48-core GPU)64 GiB · Unified memory48.3 GiB48.0 GiBOver capacity
Mac Studio M1 Ultra · 128 GB (48-core GPU)128 GiB · Unified memory48.3 GiB96.0 GiBRoom to spare
Mac Studio M1 Ultra · 64 GB (64-core GPU)64 GiB · Unified memory48.3 GiB48.0 GiBOver capacity
Mac Studio M1 Ultra · 128 GB (64-core GPU)128 GiB · Unified memory48.3 GiB96.0 GiBRoom to spare
Mac Studio M2 Max · 32 GB (30-core GPU)32 GiB · Unified memory48.3 GiB24.0 GiBOver capacity
Mac Studio M2 Max · 64 GB (30-core GPU)64 GiB · Unified memory48.3 GiB48.0 GiBOver capacity
Mac Studio M2 Max · 32 GB (38-core GPU)32 GiB · Unified memory48.3 GiB24.0 GiBOver capacity
Mac Studio M2 Max · 64 GB (38-core GPU)64 GiB · Unified memory48.3 GiB48.0 GiBOver capacity
Mac Studio M2 Max · 96 GB (38-core GPU)96 GiB · Unified memory48.3 GiB72.0 GiBRoom to spare
Mac Studio M2 Ultra · 64 GB (60-core GPU)64 GiB · Unified memory48.3 GiB48.0 GiBOver capacity
Mac Studio M2 Ultra · 128 GB (60-core GPU)128 GiB · Unified memory48.3 GiB96.0 GiBRoom to spare
Mac Studio M2 Ultra · 192 GB (60-core GPU)192 GiB · Unified memory48.3 GiB144.0 GiBRoom to spare
Mac Studio M2 Ultra · 64 GB (76-core GPU)64 GiB · Unified memory48.3 GiB48.0 GiBOver capacity
Mac Studio M2 Ultra · 128 GB (76-core GPU)128 GiB · Unified memory48.3 GiB96.0 GiBRoom to spare
Mac Studio M2 Ultra · 192 GB (76-core GPU)192 GiB · Unified memory48.3 GiB144.0 GiBRoom to spare
MacBook Pro M3 Pro · 18 GB (18-core GPU)18 GiB · Unified memory48.3 GiB13.5 GiBOver capacity
MacBook Pro M3 Pro · 36 GB (18-core GPU)36 GiB · Unified memory48.3 GiB27.0 GiBOver capacity
MacBook Pro M3 Max · 36 GB (30-core GPU)36 GiB · Unified memory48.3 GiB27.0 GiBOver capacity
MacBook Pro M3 Max · 96 GB (30-core GPU)96 GiB · Unified memory48.3 GiB72.0 GiBRoom to spare
MacBook Pro M3 Max · 48 GB (40-core GPU)48 GiB · Unified memory48.3 GiB36.0 GiBOver capacity
MacBook Pro M3 Max · 64 GB (40-core GPU)64 GiB · Unified memory48.3 GiB48.0 GiBOver capacity
MacBook Pro M3 Max · 128 GB (40-core GPU)128 GiB · Unified memory48.3 GiB96.0 GiBRoom to spare
Mac mini M4 · 16 GB (10-core GPU)16 GiB · Unified memory48.3 GiB12.0 GiBOver capacity
Mac mini M4 · 24 GB (10-core GPU)24 GiB · Unified memory48.3 GiB18.0 GiBOver capacity
Mac mini M4 · 32 GB (10-core GPU)32 GiB · Unified memory48.3 GiB24.0 GiBOver capacity
Mac mini M4 Pro · 24 GB (16-core GPU)24 GiB · Unified memory48.3 GiB18.0 GiBOver capacity
Mac mini M4 Pro · 48 GB (16-core GPU)48 GiB · Unified memory48.3 GiB36.0 GiBOver capacity
Mac mini M4 Pro · 64 GB (16-core GPU)64 GiB · Unified memory48.3 GiB48.0 GiBOver capacity
Mac mini M4 Pro · 24 GB (20-core GPU)24 GiB · Unified memory48.3 GiB18.0 GiBOver capacity
Mac mini M4 Pro · 48 GB (20-core GPU)48 GiB · Unified memory48.3 GiB36.0 GiBOver capacity
Mac mini M4 Pro · 64 GB (20-core GPU)64 GiB · Unified memory48.3 GiB48.0 GiBOver capacity
Mac Studio M4 Max · 36 GB (32-core GPU)36 GiB · Unified memory48.3 GiB27.0 GiBOver capacity
Mac Studio M4 Max · 48 GB (40-core GPU)48 GiB · Unified memory48.3 GiB36.0 GiBOver capacity
Mac Studio M4 Max · 64 GB (40-core GPU)64 GiB · Unified memory48.3 GiB48.0 GiBOver capacity
Mac Studio M4 Max · 128 GB (40-core GPU)128 GiB · Unified memory48.3 GiB96.0 GiBRoom to spare
Mac Studio M3 Ultra · 96 GB (60-core GPU)96 GiB · Unified memory48.3 GiB72.0 GiBRoom to spare
Mac Studio M3 Ultra · 256 GB (60-core GPU)256 GiB · Unified memory48.3 GiB192.0 GiBRoom to spare
Mac Studio M3 Ultra · 96 GB (80-core GPU)96 GiB · Unified memory48.3 GiB72.0 GiBRoom to spare
Mac Studio M3 Ultra · 256 GB (80-core GPU)256 GiB · Unified memory48.3 GiB192.0 GiBRoom to spare
Mac Studio M3 Ultra · 512 GB (80-core GPU)512 GiB · Unified memory48.3 GiB384.0 GiBRoom to spare
MacBook Pro M5 · 16 GB (10-core GPU)16 GiB · Unified memory48.3 GiB12.0 GiBOver capacity
MacBook Pro M5 · 24 GB (10-core GPU)24 GiB · Unified memory48.3 GiB18.0 GiBOver capacity
MacBook Pro M5 · 32 GB (10-core GPU)32 GiB · Unified memory48.3 GiB24.0 GiBOver capacity
MacBook Pro M5 Pro · 24 GB (16-core GPU)24 GiB · Unified memory48.3 GiB18.0 GiBOver capacity
MacBook Pro M5 Pro · 48 GB (16-core GPU)48 GiB · Unified memory48.3 GiB36.0 GiBOver capacity
MacBook Pro M5 Pro · 24 GB (20-core GPU)24 GiB · Unified memory48.3 GiB18.0 GiBOver capacity
MacBook Pro M5 Pro · 48 GB (20-core GPU)48 GiB · Unified memory48.3 GiB36.0 GiBOver capacity
MacBook Pro M5 Pro · 64 GB (20-core GPU)64 GiB · Unified memory48.3 GiB48.0 GiBOver capacity
MacBook Pro M5 Max · 36 GB (32-core GPU)36 GiB · Unified memory48.3 GiB27.0 GiBOver capacity
MacBook Pro M5 Max · 48 GB (40-core GPU)48 GiB · Unified memory48.3 GiB36.0 GiBOver capacity
MacBook Pro M5 Max · 64 GB (40-core GPU)64 GiB · Unified memory48.3 GiB48.0 GiBOver capacity
MacBook Pro M5 Max · 128 GB (40-core GPU)128 GiB · Unified memory48.3 GiB96.0 GiBRoom to spare
Ryzen AI Max+ 395 · 32 GB32 GiB · Unified memory48.3 GiB24.0 GiBOver capacity
Ryzen AI Max+ 395 · 64 GB64 GiB · Unified memory48.3 GiB48.0 GiBOver capacity
Ryzen AI Max+ 395 · 128 GB128 GiB · Unified memory48.3 GiB96.0 GiBRoom to spare
Ryzen AI Max+ 392 · 32 GB32 GiB · Unified memory48.3 GiB24.0 GiBOver capacity
Ryzen AI Max+ 392 · 64 GB64 GiB · Unified memory48.3 GiB48.0 GiBOver capacity
Ryzen AI Max+ 392 · 128 GB128 GiB · Unified memory48.3 GiB96.0 GiBRoom to spare
Ryzen AI Max+ 388 · 32 GB32 GiB · Unified memory48.3 GiB24.0 GiBOver capacity
Ryzen AI Max+ 388 · 64 GB64 GiB · Unified memory48.3 GiB48.0 GiBOver capacity
Ryzen AI Max+ 388 · 128 GB128 GiB · Unified memory48.3 GiB96.0 GiBRoom to spare
NVIDIA DGX Spark · 128 GB128 GiB · Unified memory48.3 GiB96.0 GiBRoom to spare
2 × GeForce RTX 309048 GiB · Multi-GPU48.3 GiB46.0 GiBConditional · Over capacity
4 × GeForce RTX 309096 GiB · Multi-GPU48.8 GiB92.0 GiBConditional · Room to spare
2 × GeForce RTX 409048 GiB · Multi-GPU48.3 GiB46.0 GiBConditional · Over capacity
2 × GeForce RTX 509064 GiB · Multi-GPU48.3 GiB62.0 GiBConditional · Room to spare
4 × GeForce RTX 5090128 GiB · Multi-GPU48.8 GiB124.0 GiBConditional · Room to spare
2 × GeForce RTX 5060 Ti 16 GB32 GiB · Multi-GPU48.3 GiB30.0 GiBConditional · Over capacity
2 × RTX A600096 GiB · Multi-GPU48.3 GiB94.0 GiBConditional · Room to spare
2 × RTX PRO 5000 Blackwell 72 GB144 GiB · Multi-GPU48.3 GiB142.0 GiBConditional · Room to spare
2 × RTX PRO 6000 Blackwell Max-Q192 GiB · Multi-GPU48.3 GiB190.0 GiBConditional · Room to spare
4 × RTX PRO 6000 Blackwell Max-Q384 GiB · Multi-GPU48.8 GiB380.0 GiBConditional · Room to spare
2 × Radeon AI PRO R970064 GiB · Multi-GPU48.3 GiB62.0 GiBConditional · Room to spare
4 × Radeon AI PRO R9700128 GiB · Multi-GPU48.8 GiB124.0 GiBConditional · Room to spare
2 × Arc Pro B7064 GiB · Multi-GPU48.3 GiB62.0 GiBConditional · Room to spare
4 × Arc Pro B70128 GiB · Multi-GPU48.8 GiB124.0 GiBConditional · Room to spare
2 × Instinct MI210 PCIe128 GiB · Multi-GPU48.3 GiB124.0 GiBConditional · Room to spare
2 × H200 NVL 141 GB PCIe282 GiB · Multi-GPU48.3 GiB278.0 GiBConditional · Room to spare
4 × H200 NVL 141 GB PCIe564 GiB · Multi-GPU48.8 GiB556.0 GiBConditional · Room to spare
GeForce RTX 3050 6 GB6 GiB · GPU48.3 GiB5.0 GiBOver capacity
GeForce RTX 3050 8 GB8 GiB · GPU48.3 GiB7.0 GiBOver capacity
GeForce RTX 3060 Ti 8 GB GDDR68 GiB · GPU48.3 GiB7.0 GiBOver capacity
GeForce RTX 30708 GiB · GPU48.3 GiB7.0 GiBOver capacity
GeForce RTX 3070 Ti8 GiB · GPU48.3 GiB7.0 GiBOver capacity
GeForce RTX 3080 10 GB10 GiB · GPU48.3 GiB9.0 GiBOver capacity
GeForce RTX 3080 12 GB12 GiB · GPU48.3 GiB11.0 GiBOver capacity
GeForce RTX 3080 Ti12 GiB · GPU48.3 GiB11.0 GiBOver capacity
Radeon RX 66008 GiB · GPU48.3 GiB7.0 GiBOver capacity
Radeon RX 6600 XT8 GiB · GPU48.3 GiB7.0 GiBOver capacity
Radeon RX 6650 XT8 GiB · GPU48.3 GiB7.0 GiBOver capacity
Radeon RX 6700 XT12 GiB · GPU48.3 GiB11.0 GiBOver capacity
Radeon RX 6750 XT12 GiB · GPU48.3 GiB11.0 GiBOver capacity
Arc A750 8 GB8 GiB · GPU48.3 GiB7.0 GiBOver capacity
Arc A770 8 GB8 GiB · GPU48.3 GiB7.0 GiBOver capacity

Architecture and calculation

Gated DeltaNet states use FP32; convolution states use the model dtype. Full-attention layers alone grow with text length. MTP speculation is excluded from cache.

  • 12 full-attention layers × 2 KV heads × (256 key + 256 value dimensions).

36 recurrent layers maintain a fixed state per request, using FP32 state and 16-bit convolution buffers. They do not allocate a full attention cache for every token.

Precision is a storage estimate. Quantized files include scales and some higher-precision tensors; use the actual file size when known. A format also needs backend and hardware support.

License: Apache 2.0. See the publisher’s terms.