Benchmark result

qwen3.8-flash-next-coder-iq1_m on NVIDIA GeForce RTX 5090 — 172.3 tok/s

Measured with Strata Strata (manual) on October 26, 2026.

What the model built

The coding scenario asks for a playable game in a single HTML file. This is exactly what the model returned, unedited; the frame only adds a line that tells this page how tall it is.

How this run compares

qwen3.8-flash-next-coder-iq1_m on other hardware

85.1tok/s86.9NVIDIA GeForce RTX 3080 +1 more
65.6tok/s72.2AMD Radeon RX 7900 XTX
62.0tok/s86.4AMD Radeon RX 9070/9070 XT/9070 GRE
43.3tok/s84.0AMD Radeon RX 9070/9070 XT/9070 GRE
42.2tok/s63.8AMD Radeon RX 7800 XT

Reproduce this run

toolStrata Strata (manual)

modelqwen3.8-flash-next-coder-iq1_m

context131,072

effortxhigh (model default)

samplingtemperature 1 · top_p 0.95 · top_k 20

clientv0.4.98+102

Set those in Strata, then:

llm-benchmark benchmark --model "qwen3.8-flash-next-coder-iq1_m" --tool "Strata"

The client prompts for context length and thinking mode, and for the KV cache dtype on Unsloth Studio. Sampling is left at the model default — the values above are what the tool reported using, not overrides the client sent.