Packets and recipes

The deployment variants of this family, at every maturity.

qwen3-14b-instruct-f723b4eB70-measured narrow baseline

Qwen3 14B Instruct Q4_K_M · TP1

Measured evidence: 38.25 tok/s on 1× B70 · rapid research snapshot

12 unique cold prompts, 128 output tokens, median tokens 1-100 after TTFT, server and request prompt caches disabled lab report — documents the result; not a step-by-step install guide

Evidence D of A–D

decodeTTFTTP1cache-zero strict suite

What has been classified

Dense scorecards summarize every declared combination; measured slices retain exact evidence links.

quant × TP

Exact measured revision and rapid runtime; TP2/4 are unmeasured gaps. Fixed: revision=qwen3-14b-instruct-f723b4e · runtime=llama.cpp SYCL fdb1db877 · mtp=0.

✓ Measured1 card, Q4_K_M TP1·Q4_K_M38.25 tok/sreport · evidence

2 untested combinations

TP2·Q4_K_MTP4·Q4_K_M

Full matrix and exact selectors

Measured results

Every number links to its proof.

family data

Measured speed (quick lab check)

Q4_K_M · one B70 · f16 KV · ctx=2048 configured; this one point is not a context curve

Q4_K_M · tensor parallel cards 138.25tok/sDecode240.04msTTFTstrict rapid snapshot · promoted conservative standalone row · evidence

Q4_K_Mmissing

evidence 1

Fine print

Qwen3 dense 14B. one measured Qwen3 14B Instruct GGUF revision. Measurements, artifact hashes, outputs, quality decisions, and speed stay pinned to their exact recorded identity.

Many people at once Not measured

Multi-user (aggregate) throughput has not been measured for this family. The lab’s one measured sweep so far is in the multi-user report; the projection block below includes a projected users curve where a projection exists.

How much faster could this get? No projection

No projection is shown: this model is not in the ML Bottleneck catalog yet. The measured numbers above stand on their own.