Radeon RX 6900 XT vs Radeon RX 9070 XT: specs and benchmark comparison
Short verdict
AMD Radeon RX 9070 XT leads in every scenario with comparable benchmarks: rendering and compute. This does not imply an advantage in tasks without benchmarks.
Compare with AIRendering and compute
B +25.6%AMD Radeon RX 9070 XT leads in this scenario (up to 25.6% in comparable benchmarks).
No comparable benchmarks yet: gaming: rasterization, gaming: ray tracing, power efficiency.
Advantages
Only measured and rated values: a noticeable benchmark gap and specs where more is better.
Where Radeon RX 6900 XT is better2
Compute units 16 more80 CU vs 64 CU
Board power 4 W lower300 W (TBP) vs 304 W (TBP)
Where Radeon RX 9070 XT is better4
Blender Benchmark 4.5.0 25.6% faster3,280 samples/min vs 2,611 samples/min
Memory speed 4 Gbps more20 Gbps vs 16 Gbps
Memory bandwidth 128 GB/s more640 GB/s vs 512 GB/s
Launch price $400 lower$599 vs $999
Benchmarks
Results are compared within one benchmark, one version and the same conditions.
- B ahead by 25.6%
Blender Benchmark 4.5.0
GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) · samples/min, higher is better
Radeon RX 6900 XT2,610.9 samples/minRadeon RX 9070 XT3,280.3 samples/minPosition among tested models (13, HIP)
A · 2,611B · 3,28002,0004,0006,000Each dot is one model’s median in this benchmark. The scale starts at zero, samples/min. AMD Radeon RX 9070 XT is ahead by 25.6% in this benchmark.
About the measurement and source
- A · AMD Radeon RX 6900 XT
2,610.9 samples/min · community measurement, sample 47 · November 14, 2026
Blender Open Data (Blender Foundation)
Median of 47 submissions (9 outliers excluded), interquartile range 2592.36–2625.33; submissions 2025-07-21 to 2026-03-05; HIP backend. Snapshot 2026-11-14.
- B · AMD Radeon RX 9070 XT
3,280.3 samples/min · community measurement, sample 666 · November 14, 2026
Blender Open Data (Blender Foundation)
Median of 666 submissions (10 outliers excluded), interquartile range 3154.63–3383.79; submissions 2025-07-17 to 2026-08-23; HIP backend. Snapshot 2026-11-14.
difference = |A − B| ÷ the smaller value × 100%. The 5% noticeable threshold is a UX heuristic, not statistical significance.
Cycles render in Blender Benchmark. Median of the sum of samples per minute over three scenes after removing outliers.
Other comparable benchmarks (1)
| Benchmark | A | B | Difference |
|---|---|---|---|
| Blender Benchmark 4.4.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 2,549 | 3,105.7 | B +21.8% |
Each benchmark version has its own scale: numbers are compared only within a row.
Reference results (17)
| Benchmark | A | B | Difference |
|---|---|---|---|
| Blender Benchmark 4.2.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 2,778.6 | — | no pair |
| Blender Benchmark 5.1.1GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 3,137.1 | 3,580.9 | reference |
| Blender Benchmark 5.2.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 2,846.4 | 3,383 | reference |
| Blender Benchmark 3.0.1GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 2,263 | — | no pair |
| Blender Benchmark 3.1.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 2,179.8 | — | no pair |
| Blender Benchmark 3.2.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 2,470.3 | — | no pair |
| Blender Benchmark 3.2.1GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 2,472.4 | — | no pair |
| Blender Benchmark 3.3.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 2,059.9 | — | no pair |
| Blender Benchmark 3.4.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 2,667.5 | — | no pair |
| Blender Benchmark 3.5.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 2,695.4 | — | no pair |
| Blender Benchmark 3.6.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 2,646.5 | — | no pair |
| Blender Benchmark 4.0.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 2,624.2 | — | no pair |
| Blender Benchmark 4.1.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 2,811.3 | — | no pair |
| Blender Benchmark 4.3.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 2,793.3 | 3,323.5 | reference |
| Blender Benchmark 5.0.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 2,631.3 | 2,911 | reference |
| Blender Benchmark 5.1.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 3,154.3 | 3,624.4 | reference |
| Blender Benchmark 5.3.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | — | 3,179.3 | no pair |
Each benchmark version has its own scale: numbers are compared only within a row.
Specifications
| Specification | A · Radeon RX 6900 XT | B · Radeon RX 9070 XT | Difference |
|---|---|---|---|
| Chip | |||
| Architecture | RDNA 2 | RDNA 4 | differs |
| Compute unitsSM (NVIDIA), CU (AMD) and Xe-core (Intel) are different units and are not compared numerically across vendors. | 80 CU | 64 CU | A: 16 more |
| Shader unitsShader unit counts are comparable only within one architecture. | 5,120 (Stream processors) | 4,096 (Stream processors) | not directly comparable |
| Clocks | |||
| Game Clock (AMD) | 2,015 MHz | 2,400 MHz | B: 385 MHz more |
| Boost clockClocks of different architectures do not reflect performance directly. For laptop chips, the clock depends on the laptop’s power limit. | 2,250 MHz | 2,970 MHz | B: 720 MHz more |
| Memory | |||
| Video memory | 16 GB | 16 GB | equal |
| Memory type | GDDR6 | GDDR6 | equal |
| Memory bus | 256 bit | 256 bit | equal |
| Memory speed | 16 Gbps | 20 Gbps | B: 4 Gbps more |
| Memory bandwidthExcludes cache (Infinity Cache / L2), which affects effective bandwidth differently. | 512 GB/s | 640 GB/s | B: 128 GB/s more |
| Power and dimensions | |||
| Board powerTGP (NVIDIA) and TBP (AMD/Intel) are calculated differently; higher is not better. For laptops, the chip maker sets the range and the laptop maker picks the value, so the same chip behaves differently in different laptops. | 300 W (TBP) | 304 W (TBP) | B: 4 W higher |
| Recommended PSU | 850 W | 750 W | A: 100 W higher |
| Interface | no data | PCIe 5.0 x16 | no data for A |
| Length (reference / Founders Edition)Partner cards have different dimensions; check the specific SKU. | 267 mm | no data | no data for B |
| Thickness (slots) | 2.5 | no data | no data for B |
| Power connectors (reference) | 2x 8-pin (reference card) | no data | no data for B |
Chip
Architecture
ARDNA 2
BRDNA 4
differsCompute units
A80 CU
B64 CU
A: 16 moreSM (NVIDIA), CU (AMD) and Xe-core (Intel) are different units and are not compared numerically across vendors.
Shader units
A5,120 (Stream processors)
B4,096 (Stream processors)
not directly comparableShader unit counts are comparable only within one architecture.
Clocks
Game Clock (AMD)
A2,015 MHz
B2,400 MHz
B: 385 MHz moreBoost clock
A2,250 MHz
B2,970 MHz
B: 720 MHz moreClocks of different architectures do not reflect performance directly. For laptop chips, the clock depends on the laptop’s power limit.
Memory
Video memory
A16 GB
B16 GB
equalMemory type
AGDDR6
BGDDR6
equalMemory bus
A256 bit
B256 bit
equalMemory speed
A16 Gbps
B20 Gbps
B: 4 Gbps moreMemory bandwidth
A512 GB/s
B640 GB/s
B: 128 GB/s moreExcludes cache (Infinity Cache / L2), which affects effective bandwidth differently.
Power and dimensions
Board power
A300 W (TBP)
B304 W (TBP)
B: 4 W higherTGP (NVIDIA) and TBP (AMD/Intel) are calculated differently; higher is not better. For laptops, the chip maker sets the range and the laptop maker picks the value, so the same chip behaves differently in different laptops.
Recommended PSU
A850 W
B750 W
A: 100 W higherInterface
Ano data
BPCIe 5.0 x16
no data for ALength (reference / Founders Edition)
A267 mm
Bno data
no data for BPartner cards have different dimensions; check the specific SKU.
Thickness (slots)
A2.5
Bno data
no data for BPower connectors (reference)
A2x 8-pin (reference card)
Bno data
no data for B
“More” does not always mean “better”: higher power draw or size is a system requirement, not an advantage (such bars are dimmed).
Compare with AI
An AI model reviews the specs and benchmarks of this pair and gives a plain-language verdict on who each model suits.
Questions and answers
How much video memory do AMD Radeon RX 6900 XT and AMD Radeon RX 9070 XT have?
AMD Radeon RX 6900 XT: 16 GB GDDR6, 256-bit bus. AMD Radeon RX 9070 XT: 16 GB GDDR6, 256-bit bus.
What power supply does the manufacturer recommend?
For AMD Radeon RX 6900 XT, 850 W or more; for AMD Radeon RX 9070 XT, 750 W or more. This is a recommendation for a typical system; also check the power connectors of the specific card model.
Which is faster: AMD Radeon RX 6900 XT or AMD Radeon RX 9070 XT?
AMD Radeon RX 9070 XT leads in every scenario with comparable benchmarks: rendering and compute. This does not imply an advantage in tasks without benchmarks.
See also
Data sources (3)
- AMD — specifications of AMD Radeon RX 6900 XT, specifications of AMD Radeon RX 9070 XTchecked November 7, 2026
- Blender Open Data (Blender Foundation) — benchmark Blender Benchmark 4.5.0, benchmark Blender Benchmark 5.1.1, benchmark Blender Benchmark 5.2.0, benchmark Blender Benchmark 4.3.0, benchmark Blender Benchmark 4.4.0, benchmark Blender Benchmark 5.0.0, benchmark Blender Benchmark 5.1.0, benchmark Blender Benchmark 5.3.0, benchmark Blender Benchmark 4.2.0, benchmark Blender Benchmark 3.0.1, benchmark Blender Benchmark 3.1.0, benchmark Blender Benchmark 3.2.0, benchmark Blender Benchmark 3.2.1, benchmark Blender Benchmark 3.3.0, benchmark Blender Benchmark 3.4.0, benchmark Blender Benchmark 3.5.0, benchmark Blender Benchmark 3.6.0, benchmark Blender Benchmark 4.0.0, benchmark Blender Benchmark 4.1.0checked November 14, 2026
- Wikipedia (English) — release date and launch pricechecked November 14, 2026
How we compare: methodology. Where the data comes from: sources.