llamaperf

AMD RX 9070 XT 16GB reports

27 reports. These are individual setups, not a controlled benchmark. AMD RX 9070 XT 16GB speeds and models →

Speeds are as each source reported them. GPU count, offloading and concurrent requests change them, and prompt-processing speed is input throughput, not time to first token. Check the full setup before comparing two reports.

Qwen3.8 27B

AMD RX 9070 XT 16GB · llama.cpp · 131,072 ctx

AMD hardware
generation:
50.0 tokens/s
quant:
Q6 (GGUF)
kv:
Q8

User reports 50 t/s for Qwen 3.8 27B on an RX 9070 XT 16GB. Setup is llama.cpp (Vulkan build) with a Q6 quant, 131072 context, Q8 K cache and Q4 V cache, and DFlash2 speculative decoding with 7 max drafts. The rig also has an R9700 AI Pro 32GB and 64GB DDR5 6000; the user asks what speeds and flags others get, especially on Windows/AMD.

Sep 28, 2026

Qwen3.8 27B Swift-1.5

AMD RX 9070 XT 16GB · llama.cpp · 65,536 ctx

Tone: positiveAMD hardware
generation:
43.7 tokens/s
prompt processing (prefill):
700.8 tokens/s
quant:
IQ3_S (GGUF)
kv:
q5_0-q4_1
tool-uselong-context

User reports Qwen3.8 27B (Swift-1.5-GSQ-RCO IQ3_S with MTP) at 43.7 tok/s decode and 700.8 t/s prefill on an RX 9070 XT 16GB. Setup is llama.cpp (server-vulkan b11176) with IQ3_S weights and a q5_0-q4_1 KV cache at 65536 context, one card. Decode is the mean of 3 runs of a fixed 512-token generation; prefill is one cold 30,461-token prompt. At 131072 context the same Vulkan launch gives 37.4 tok/s decode and 701.6 t/s prefill, while the pinned ROCm image (b10884) cannot load at 128K and decodes 33.5 tok/s at 65536. The user attributes the ROCm failure to a missing FlashAttention vector kernel for the q5_0/q4_1 K/V pair, which promotes both to f16 and exhausts the 16GB card.

Sep 27, 2026

Qwen3.8 27B

2× AMD RX 9070 XT 16GB · llama.cpp · 98,304 ctx

AMD hardware
generation:
24.7 tokens/s
prompt processing (prefill):
2061.1 tokens/s
quant:
Q4_K_M (GGUF)
kv:
f8_e4m3
codingagentictool-usevisionlong-context

User reports Qwen3.8-27B at 2061.11 t/s prompt and 24.70 t/s decode on two Radeon RX 9070 XT 16 GB cards under ROCm 7.2 on Windows. Setup is a llama.cpp fork with Q4_K_M GGUF, f8_e4m3 KV cache, FlashAttention, batch 8192 / ubatch 1024, layer split across both GPUs, one server slot, and cold prompt processing at 98,304 context (L3 lane). With MTP n3 the same L1 lane gives 1817.81 t/s prompt and 36.73 t/s decode at 46.8% acceptance. Vulkan rows are lower on prompt processing (L1 1569.17, L2 1470.80, L3 1243.99 t/s) and lead only the L1 decode rows. Linux ROCm 10 results reach 2253.06 t/s prompt with MXFP4-requant and 50.13 t/s decode with Q4_K_M plus MTP n3.

Sep 27, 2026
AMD hardware
prompt processing (prefill):
706.0 tokens/s
quant:
Q4_K_M (GGUF)

User reports Qwen3.6-35B-A3B at 706.0 t/s prompt processing on an AMD Radeon RX 9070 XT 16GB under ROCm. Setup is llama.cpp with Q4_K_M GGUF, flash attention enabled, -ncmoe 40 and -ub 512, on a Ryzen 9 5950X with 64 GB DDR4-3200. The same configuration under Vulkan reached 333.4 t/s prompt and 29.3 t/s generation; the user retired Vulkan in favour of ROCm.

Sep 27, 2026

Qwen3.5 9B

AMD RX 9070 XT 16GB · llama.cpp · 65,536 ctx

Tone: positiveAMD hardware
generation:
62.0 tokens/s
quant:
Q6_K_XL (GGUF)

User benchmarks Qwen3.5-9B at 62 t/s on an AMD Radeon RX 9070 XT (RDNA4/gfx1201) using llama-server with the Vulkan backend. Setup is llama.cpp with GGUF Q6_K_XL weights at 65536 tokens of context, single request, no speculation. The same model under vLLM ROCm 7.2 with FP8 weights reached 48 t/s, which the user attributes to vLLM lacking native gfx1201 kernel support and falling back to FP32 dequantization. The user reports llama-server Vulkan is 29% faster on this hardware and recommends it over vLLM until RDNA4 support lands upstream.

Sep 27, 2026

Qwopus3.5 9B Coder

AMD RX 9070 XT 16GB · 128,000 ctx

AMD hardware
generation:
150.0 tokens/s
quant:
Q8_0 (GGUF)
kv:
F16
coding

User compares Qwopus3.5-9B-Coder at Q8_0 against Qwen3.8-27B at IQ4_XS. The 9B runs at 150+ t/s with 128K context and a BF16 KV cache. The 27B is limited to ~80K context with Kvarn4 and MTP with a Q4 cache at 20 t/s. User asks whether the 9B is still worth using for coding in 2026.

Sep 23, 2026

Qwen3.8 27B Swift

2× AMD RX 9070 XT 16GB · llama.cpp · 131,072 ctx

Tone: mixedAMD hardware
generation:
50.0 tokens/s
quant:
Q6 (GGUF)
kv:
Q8

User reports Qwen 3.8 Swift at a ceiling of 50 t/s on an RX 9070 XT 16GB and a Radeon AI PRO R9700 32GB. Setup is llama.cpp with the Vulkan backend, a Q6 quant, 131K context, and Q8 KV cache for both K and V, with a DFlash2 drafter. User asks whether others get more than 50 t/s on AMD 9070 XT or R9700 hardware with a usable context size, and whether a different OS could add 30+ t/s.

Sep 21, 2026