Pick your setup

Models

Before you tune anything, you pick a model — and that choice sets the quality and capability ceiling. Active parameters set an important speed baseline, but quantization, runtime, card count, graph mode, and speculation can move the observed rate substantially.

The short version

  • Total parameters decide how much VRAM the model needs. Active parameters strongly influence decode speed.
  • Dense models use all their weights per token. Mixture-of-Experts (MoE) models use only a slice — so they're fast for their size.
  • A name like A4B or A3B tells you the active size (≈4B or 3B active) — that's your speed hint.
  • Fit first, then quality/capability and packet evidence; compare decode, prefill, TTFT, memory, and context only inside that viable set.

Dense vs mixture-of-experts

A dense model runs every one of its weights for every token. A 27B dense model does ~27B parameters' worth of work per token — accurate, but heavy on the memory reads that set decode speed.

A mixture-of-experts model is split into many "experts," and a router picks only a few per token. It might hold 30B parameters total but activate just 3–4B on any given token. You pay the VRAM of the big model but the speed of a small one. That's the trick behind the fastest models on modest hardware.

Same card, two designs

Bar chart: Gemma-4-26B A4B mixture-of-experts at about 125 tokens per second versus Qwen3.8-27B at about 28 tokens per second, both on one B70.
Both on one B70. Gemma-4-26B is a sparse MoE (~4B active); Qwen3.8-27B activates far more per token. Decode tracks active size — total parameters (26B vs 27B) barely differ. (Quant also differs, Q8 vs Q4; the design gap is the dominant factor.) Lab-measured
Reading a model name

Qwen3.8-27B is a dense hybrid model: it has no A3B suffix and does not get MoE-style 3B-active decode. Gemma-4-26B-A4B is about 4B active; Nemotron-30B-A3B is about 3B active. When a model really has an A#B suffix, that active size is the better speed hint.

The VRAM math

Whatever the design, the model's weights must fit in memory, and you need room left for the KV cache (your context) on top. A rough single-card budget on a 32 GiB B70:

  • Weights: total params × bits-per-weight ÷ 8. A 27B model at 4-bit (Q4) ≈ 15–18 GB; at 8-bit (Q8) ≈ 27–30 GB. See Quantization.
  • KV cache: grows with context; can be several GB to 16 GB at very long context.
  • Rule of thumb: keep weights + expected KV under ~30 GB on a 32 GiB card. If a Q8 model is tight, drop to Q4 or add a card.

Selected lab measurements

What we've actually measured on B70s. Rates are for different card counts and methods — read the columns, don't compare rows blindly. Lab-measured

ModelDesignCardsDecode tok/sNote
Gemma-4-26B A4BMoE (~4B active)1124.98Runs on one card; measured
Qwen3.8-27Bhybrid, Q4_K_M127.82Quality-checked; speed measured out to 32K
Qwen3.8-27BQ8_0236.778-bit: slower, more conservative
Qwen3.8-27BINT4 + MTP 52101.2Research result (not yet repeatable run to run)
Laguna S 2.1INT4 + DFlash4125.46Lab record run
Muse-Glimmer-30BQ8/WOQ + DFlash4100.37Average across repeat runs

The Qwen 27B family page keeps weights, quantizations, TP, MTP, context, prefill, TTFT, quality, and gaps together. The full filterable packet list lives in the Recipes.

How to choose

  1. Match to your cards. One B70? A fast MoE (Gemma-4-26B) or a Q4 27B. More cards unlock bigger models and speculative records.
  2. Match to your job. Coding and agents need tool behavior and responsiveness; document work needs long context; image tasks need a vision-enabled packet such as the Qwen3.8 Q5_K_S flagship.
  3. Then optimize. Once the model fits and runs, the other nine guides are how you make it faster.
What to be aware of

A high headline number can hide caveats: it may be a research lane with unresolved determinism, a narrow benchmark shape, or a different card count than yours. It may also be a model that's fast but weaker at your task. We now have B70 measurements for Ornith, Nemotron, and LFM2.5, but their packets still carry different maturity and quality boundaries. Always check the evidence label and exact configuration behind a rate.

Recommendation

For most people on a single B70: start with Gemma-4-26B for a fast multimodal lane, Qwen3.8-27B Q4_K_M for the validated text package and measured 32K curve, or the separate Q5_K_S 256K + vision packet when those capabilities matter more than decode. Get one exact packet running, then tune.