AMD · Radeon RX 9000 Series
AMD Radeon RX 9070 XT
AMD Radeon RX 9070 XT specs, benchmarks, and best prices. 16GB GDDR6, 4096 stream processors, RDNA 4. Competes with RTX 5070 Ti at $599.
Performance
65
BGaming
67
BValue
72
BTypical street price
$600–750
Launch MSRP $599 · Released Mar 2025
Prices vary with the 2026 memory shortage. As an Amazon Associate we earn from qualifying purchases.
How This Compares
Price vs gaming performance among similar GPUs (within ±$200 MSRP). Higher and further left = better value.
Full Specifications
| VRAM | 16GB |
| Memory Type | GDDR6 |
| Boost Clock | 2970MHz |
| Base Clock | 2400MHz |
| Shader/Stream Processors | 4096 |
| Performance | 48.7TFLOPS |
| TDP | 304W |
| Memory Bus | 256-bit |
| Bandwidth | 640GB/s |
| Architecture | RDNA 4 |
| Process Node | 4nm |
| RT Cores | 64 |
| AI Upscaling | FSR 4 |
| Interface | PCIe 4.0 x16 |
| Power Connector | 2x 8-pin |
| Card Length | 312mm |
| Release Date | 2025-03-06 |
Local AI on the RX 9070 XT
With 16GB of VRAM, here's what this card can run and how fast — estimated from memory bandwidth, the number that actually decides LLM speed.
| Model | Best quant that fits | Est. speed | Verdict |
|---|---|---|---|
| gpt-oss-20b | Q4_K_M | ~100 tok/s (Instant-feeling) | Runs great |
| Qwen3 8B | Q8_0 | ~37 tok/s (Fast) | Runs great |
| Llama 3.1 8B | Q8_0 | ~38 tok/s (Fast) | Runs great |
| Qwen3.5 35B-A3B | — | — | Partial offload |
| Qwen3 32B | — | — | Partial offload |
| DeepSeek R1 Distill Qwen 32B | — | — | Partial offload |
| Gemma 4 31B | — | — | Partial offload |
| Qwen3 30B-A3B | — | — | Partial offload |
3 of 8 featured models run fully in VRAM on this card. Check a custom model size →
Compare RX 9070 XT With
Understand the Specs
Browse all terms →VRAM (Video RAM)
Dedicated memory on a graphics card used to store textures, frame buffers, and render data. More VRAM allows higher resolutions and more detailed textures.
GPU (Graphics Processing Unit)
A specialized processor designed for parallel computation, primarily used for rendering graphics but increasingly important for AI and compute workloads.
Ray Tracing
A rendering technique that simulates how light physically behaves — bouncing, reflecting, and refracting — to produce photorealistic lighting, shadows, and reflections in games.
FSR (FidelityFX Super Resolution)
AMD's upscaling technology that boosts frame rates by rendering at lower resolution and reconstructing detail. FSR 4 uses AI on RDNA 4 hardware; older versions work on any GPU.
PCIe (PCI Express)
The high-speed interface connecting GPUs, SSDs, and expansion cards to your CPU and motherboard. Each generation doubles bandwidth per lane.
Memory Bandwidth
The rate at which data can be read from or written to memory, measured in GB/s. Critical for GPU performance and AI workloads. Determined by memory type, bus width, and clock speed.