AMD Radeon RX 6900 XT
16 GB decides what fits. 512 GB/s decides how fast it runs once it does.
Open in the calculator →Best models for 16 GB, on every card that size →
Find one: Amazon ↗ · eBay (new and used) ↗Store links may pay us a commission. They never decide which card is suggested — the memory arithmetic does.
- Largest popular model that fitsgpt-oss-20b21B params · Q4_K_M · needs 13.8 GB~159 tok/s Faster than you readOpen in the calculator →
- Best fast pickERNIE-4.5-21B-A3B-PT22B params · Q4_K_M · needs 14.7 GB~157 tok/s Faster than you readOpen in the calculator →
- Most downloaded that fitsQwen3.5-9B9.7B params · Q4_K_M · needs 7.05 GB~82 tok/s Faster than you readOpen in the calculator →
At 8,192 tokens of context with the whole model in device memory. Speeds are estimates from memory bandwidth, not benchmarks run on this card.
153 of 320 fit entirely
153 of the 320 models the engine can size fit entirely in device memory at Q4_K_M where it is published, otherwise the nearest published format, and 8,192 tokens. 73 run with some layers on system memory (32 GB assumed), and 94 do not run at all.
Featured models · Q4_K_M at 8,192 tokens, including offload
| Model | Needs | Verdict | Decode | Calculator |
|---|---|---|---|---|
| Qwen3.5-2B2.3B parameters | 2.32 GB | fits | ~323 tok/s | Open → |
| NVIDIA-Nemotron-3-Nano-4B-BF164.0B parameters | 3.46 GB | fits | ~150 tok/s | Open → |
| Qwen3-VL-8B-Instruct8.8B parameters | 7.04 GB | fits | ~68 tok/s | Open → |
| Qwen3.5-9B9.7B parameters | 7.05 GB | fits | ~82 tok/s | Open → |
| gemma-4-12B-it12B parameters | 9.54 GB | fits | ~49 tok/s | Open → |
| gpt-oss-20b21B parameters | 13.8 GB | fits | ~159 tok/s | Open → |
| Qwen3.8-27B28B parameters | 18.6 GB | 10 of 64 layers on system RAM (2.66 GB) | ~15 tok/s with offload | Open → |
| Qwen3.6-27B28B parameters | 18.6 GB | 10 of 64 layers on system RAM (2.66 GB) | ~15 tok/s with offload | Open → |
| Kimi-Linear-48B-A3B-Instruct49B parameters | 30.4 GB | 14 of 27 layers on system RAM (15.3 GB) | ~57 tok/s with offload | Open → |
| Qwen3.8-Flash-Next180B parameters | 111.6 GB | needs 96.7 GB of system RAM; 32 GB assumed | does not run | Open → |
| DeepSeek-V4-Flash-Vision-Exp305B parameters | 188.1 GB | needs 174.2 GB of system RAM; 32 GB assumed | does not run | Open → |
| GLM-5.3-Flash321B parameters | 198.3 GB | needs 184.1 GB of system RAM; 32 GB assumed | does not run | Open → |
These examples are selected from prominent labs using the catalogue’s latest Hugging Face 30-day downloads and repository-creation freshness signal, with newer releases guaranteed a place. Offloaded rows assume 32 GB of system RAM, and a speed is only shown for a row that runs. Fitting in memory is not the same as loading: whether the runtime and version you have supports each architecture and format on this machine has not been tested here. Each “Open” link carries the same model, format, context and RAM into the calculator.
Fits entirely in AMD Radeon RX 6900 XT memory — 153 of 320 sized models
The most demanding model that fits is gpt-neox-20b at Q4_K_M: 15.7 GB of the 16 GB, leaving 0.25 GB spare.
Fully resident at 8,192 tokens (or the model’s own maximum, where that is shorter), offload off, at 512 GB/s. Each row is a run of the engine for this configuration; the rows start with current, prominent releases and “fits” is memory, not a tested runtime. Older or less prominent models remain available through this search and “Show all”.
Showing 40 of 153 models that fit.
| Model | Format | Needs | Spare | Decode | Downloads / 30d | Calculator |
|---|---|---|---|---|---|---|
| Qwen3.5-9BQwen · 9.7B params | Q4_K_M | 7.05 GB | 8.95 GB | ~82 tok/s | 9M | Open → |
| Qwen3.5-4BQwen · 4.7B params | Q4_K_M | 3.98 GB | 12.0 GB | ~148 tok/s | 7.8M | Open → |
| gemma-4-12B-itGoogle · 12B params | Q4_K_M | 9.54 GB | 6.46 GB | ~49 tok/s | 1.9M | Open → |
| Qwen3.5-2BQwen · 2.3B params | Q4_K_M | 2.32 GB | 13.7 GB | ~323 tok/s | 4.9M | Open → |
| gemma-4-E4B-itGoogle · 8.0B params | Q4_K_M | 5.86 GB | 10.1 GB | ~132 tok/s | 4.4M | Open → |
| NVIDIA-Nemotron-3-Nano-4B-BF16NVIDIA · 4.0B params | Q4_K_M | 3.46 GB | 12.5 GB | ~150 tok/s | 3.4M | Open → |
| gemma-4-E2B-itGoogle · 5.1B params | Q4_K_M | 4.02 GB | 12.0 GB | ~348 tok/s | 3M | Open → |
| Qwen3.5-0.8BQwen · 873M params | Q4_K_M | 1.46 GB | 14.5 GB | ~708 tok/s | 2.6M | Open → |
| Qwen3-VL-8B-InstructQwen · 8.8B params | Q4_K_M | 7.04 GB | 8.96 GB | ~68 tok/s | 14.6M | Open → |
| North-Micro-Vision-InstructCohereLabs · 2.5B params | Q4_K_M | 2.99 GB | 13.0 GB | ~213 tok/s | 180.7K | Open → |
| granite-4.2-8bIBM · 8.8B params | Q4_K_M | 7.49 GB | 8.51 GB | ~62 tok/s | 128.4K | Open → |
| granite-4.1-3bIBM · 3.4B params | Q4_K_M | 3.72 GB | 12.3 GB | ~144 tok/s | 521.8K | Open → |
| LFM2.5-2.6BLiquidAI · 2.7B params | Q4_K_M | 2.61 GB | 13.4 GB | ~231 tok/s | 108.1K | Open → |
| Qwen3-VL-4B-InstructQwen · 4.4B params | Q4_K_M | 4.72 GB | 11.3 GB | ~108 tok/s | 3.5M | Open → |
| Qwen3-VL-2B-InstructQwen · 2.1B params | Q4_K_M | 3.05 GB | 13.0 GB | ~195 tok/s | 2.8M | Open → |
| granite-4.2-3bIBM · 3.7B params | Q4_K_M | 3.72 GB | 12.3 GB | ~144 tok/s | 50.6K | Open → |
| LFM2.5-230MLiquidAI · 230M params | Q4_K_M | 1.05 GB | 14.9 GB | ~1610 tok/s | 88.6K | Open → |
| Qwen3-0.6BQwen · 752M params | Q4_K_M | 2.22 GB | 13.8 GB | ~300 tok/s | 29.7M | Open → |
| gpt-oss-20bOpenAI · 21B params | Q4_K_M | 13.8 GB | 2.24 GB | ~159 tok/s | 6.6M | Open → |
| granite-4.1-8bIBM · 8.8B params | Q4_K_M | 7.49 GB | 8.51 GB | ~62 tok/s | 179.3K | Open → |
| LFM2.5-VL-3BLiquidAI · 3.1B params | Q4_K_M | 2.85 GB | 13.1 GB | ~231 tok/s | 26.7K | Open → |
| MiMo-V2.6-Distill-Qwen-9BNEWXiaomiMiMo · 9.4B params | Q4_K_M | 6.90 GB | 9.10 GB | ~82 tok/s | 14.6K | Open → |
| Qwen3-4B-Instruct-2507Qwen · 4.0B params | Q4_K_M | 4.72 GB | 11.3 GB | ~108 tok/s | 3.7M | Open → |
| Ling-3.0-tinyinclusionAI · 7.9B params | Q4_K_M | 5.85 GB | 10.1 GB | ~465 tok/s | 17.7K | Open → |
| Qwen3-8BQwen · 8.2B params | Q4_K_M | 7.04 GB | 8.96 GB | ~68 tok/s | 10.7M | Open → |
| Olmo-3-7B-Instructallenai · 7.3B params | Q4_K_M | 7.96 GB | 8.04 GB | ~58 tok/s | 481.7K | Open → |
| Qwen3-4BQwen · 4.0B params | Q4_K_M | 4.51 GB | 11.5 GB | ~108 tok/s | 7.8M | Open → |
| LFM2.5-8B-A1BLiquidAI · 8.5B params | Q4_K_M | 6.06 GB | 9.94 GB | ~360 tok/s | 32.4K | Open → |
| LFM2.5-350MLiquidAI · 354M params | Q4_K_M | 1.13 GB | 14.9 GB | ~1039 tok/s | 69.9K | Open → |
| Hy-MT2-1.8BTencent · 2.0B params | Q4_K_M | 2.47 GB | 13.5 GB | ~239 tok/s | 28.9K | Open → |
| LLaDA2.0-miniinclusionAI · 16B params | Q4_K_M | 11.0 GB | 5.00 GB | ~373 tok/s | 217.2K | Open → |
| LFM2.5-1.2B-InstructLiquidAI · 1.2B params | Q4_K_M | 1.63 GB | 14.4 GB | ~383 tok/s | 119.2K | Open → |
| Ministral-3-14B-Instruct-2512Mistral AI · 14B params | Q4_K_M | 10.4 GB | 5.62 GB | ~43 tok/s | 251K | Open → |
| Qwen3-1.7BQwen · 2.0B params | Q4_K_M | 3.02 GB | 13.0 GB | ~195 tok/s | 3.1M | Open → |
| Nemotron-3.5-Content-SafetyNVIDIA · 4.3B params | Q4_K_M | 3.73 GB | 12.3 GB | ~144 tok/s | 14.1K | Open → |
| Hy-MT2-7BTencent · 8.0B params | Q4_K_M | 6.50 GB | 9.50 GB | ~70 tok/s | 15.4K | Open → |
| Qwen3-14BQwen · 15B params | Q4_K_M | 11.1 GB | 4.86 GB | ~40 tok/s | 2.7M | Open → |
| GLM-4.6V-Flashzai-org · 10B params | Q4_K_M | 7.45 GB | 8.55 GB | ~70 tok/s | 103.7K | Open → |
| Qwen2.5-VL-7B-InstructQwen · 8.3B params | Q4_K_M | 6.36 GB | 9.64 GB | ~82 tok/s | 5.8M | Open → |
| SmolLM3-3BHuggingFaceTB · 3.1B params | Q4_K_M | 3.46 GB | 12.5 GB | ~159 tok/s | 617K | Open → |
Sized at the format most people actually download, not at FP16. The catalogue ordered by downloads →
Find the AMD Radeon RX 6900 XT: Amazon ↗ · eBay (new and used) ↗Store links may pay us a commission. They never decide which card is suggested — the memory arithmetic does.
The Radeon RX 6900 XT, model by model
One page per model: whether it fits this device, at which formats, and how fast.
- Qwen3-0.6B
- Qwen3-VL-8B-Instruct
- gemma-4-26B-A4B-it
- Qwen3-8B
- gemma-4-31B-it
- Qwen3.5-9B
- Qwen2.5-0.5B-Instruct
- Qwen2.5-7B-Instruct
- Qwen3.5-4B
- Qwen3-4B
- Qwen2.5-1.5B-Instruct
- Qwen3.8-27B
All 70Fewer
- gpt-oss-20b
- Qwen2.5-VL-7B-Instruct
- GLM-5.3-Flash
- Qwen3.5-2B
- dolphin-2.9.1-yi-1.5-34b
- DeepSeek-V4-Flash-0731
- gpt-oss-120b
- gemma-4-E4B-it
- Qwen2.5-3B-Instruct
- Qwen3-32B
- Qwen3-4B-Instruct-2507
- DeepSeek-V3.2
- Qwen3-VL-4B-Instruct
- pythia-160m
- NVIDIA-Nemotron-3-Nano-4B-BF16
- Qwen3.6-35B-A3B
- Qwen3-1.7B
- gemma-4-E2B-it
- OTel-2.0-LLM-31B-IT
- Qwen3-VL-2B-Instruct
- Qwen3-14B
- Qwen3.5-0.8B
- JiRackUltra_1b
- Qwen3.6-27B
- Qwen2.5-VL-3B-Instruct
- Qwen2.5-Coder-7B-Instruct
- Mistral-7B-Instruct-v0.3
- Qwen2.5-32B-Instruct
- gemma-4-12B-it
- Qwen3.5-27B
- SmolLM2-135M-Instruct
- SmolLM2-135M
- GLM-4.7-Flash
- Mistral-7B-Instruct-v0.2
- Qwen2.5-Coder-14B-Instruct
- Qwen2.5-14B-Instruct
- Qwen3.5-35B-A3B
- Qwen3-30B-A3B
- Qwen2.5-0.5B
- Ornith-1.0-35B
- pythia-70m-deduped
- DeepSeek-V3-0324
- Qwen3.8-Flash-Next
- GLM-5.3
- TinyLlama-1.1B-Chat-v1.0
- DeepSeek-V3
- Kimi-K3
- NVIDIA-Nemotron-3-Super-120B-A12B-BF16
- MiniMax-M2.7
- DeepSeek-V4-Flash
- MiniCPM5-2B
- DeepSeek-R1
- DeepSeek-R1-Distill-Qwen-1.5B
- Ornith-1.0-9B
- DeepSeek-V4-Flash-Vision-Exp
- DeepSeek-Coder-V2-Lite-Instruct
- Qwen2.5-Coder-32B-Instruct
- NVIDIA-Nemotron-3-Nano-30B-A3B-BF16
The specification behind every figure
What the manufacturer publishes for this device, and the pages it was read from.
Manufacturer specification
| Memory scope | dedicated |
|---|---|
| Capacity | 16 GB |
| Published options | 16 GB |
| Bandwidth | 512 GB/s |
| Memory type | GDDR6 |
| Bus width | 256-bit |
| FP32 peak | Not published |
| Dense matrix peak | Not published without sparsity |
| Typical Board Power | 300 W |
Source ledger
Caveats
- AMD publishes bandwidth as "Up to 512 GB/s"; that ceiling is the figure used. The Infinity Cache raises AMD's "effective" figure for games; decoding a model streams far more than the cache holds, so the raw memory bandwidth is the one that decides the speed.
- No dense matrix throughput is published for this device in a form this catalogue accepts, so no compute roof is priced and time to first token is withheld.
Published capacity is a hardware ceiling, not guaranteed free runtime memory. The calculator shows the runtime reserve separately rather than folding it into a single number.
Run the diagnostic on the AMD Radeon RX 6900 XT →