Hugging Face does not say who downloads a model or with what. These are the proxies: API providers listing it, quantized builds, fine-tunes/adapters/merges, and Spaces that load it.
Served by 2 API providers; 1 quantized build for local use.
Everything else, folded: the full tables, every source.
What the card says
2 cited sentences
Summary generated, cited
Designed for semantic search, it maps sentences and paragraphs to a 768 dimensional dense vector space.[1] The model was trained on 215M (question, answer) pairs from diverse sources.[2]
Generated 2026-10-07 15:01 UTC by parasail-gemma-4-26b-a4b-it; each numeral points at the fact it rests on. Not a verdict.
Use-case evidence
2 use cases
What it is good at 2 use cases
Labels are evidence, not verdicts: measured a third-party leaderboard row for this model; author-reported a number the model author published in the card or eval files; structural a deterministic signal from the name, config or chat template, or a decision-model estimate. Each chip links to its source. All use cases.
Hosted endpoints observed for multi-qa-mpnet-base-dot-v1, sorted by input price (USD per million tokens). Each row names its observer and date; prices change daily, so check the source before quoting. Cheapest observed input price: $0.0050/M.
Ranked by source signals (HN, Lobsters, authority domains, official posts) and recency. Summaries are generated from a stored copy of the article text; every sentence cites a span of that text, and nothing is shown that failed the citation check.
blog.google2026-05-05authorityHN front pageabout the multi familyHN 687 pts · 330 comments796 words
Google released Multi-Token Prediction (MTP) drafters for the Gemma 4 family under the same open-source Apache 2.0 license as Gemma 4. These drafters deliver up to a 3x speedup for models like the 26B MoE and 31B Dense versions without quality degradation. Benchmarks on hardware such as the NVIDIA RTX PRO 6000 show that increasing batch sizes to 4 to 8 unlocks up to a ~2.2x speedup locally.written, reviewed, cited
The MTP drafter shares the target model's KV cache to avoid recalculating context the larger model has already processed.
For E2B and E4B edge models, an efficient clustering technique in the embedder accelerates generation where logit calculation is a bottleneck.
The 26B mixture-of-experts model presents unique routing challenges at a batch size of 1 on Apple Silicon.
multi-qa-mpnet-base-dot-v1 is an open-weight sentence similarity model from Sentence Transformers designed for semantic search. It has 109M parameters and a 514-token context, requiring about 0.3 GB of GPU memory at 16-bit precision. The model was trained on 215M question and answer pairs, but the license field is empty, so commercial use requires confirmed publisher terms. At 16-bit, it runs on a 1x MI300X for $1.85 an hour.written, reviewed, cited
The model requires chunking documents first because its 514-token input limit is too short for long contexts.
The total download size is 3.8 GB because the repository includes safetensors, onnx, openvino, and pytorch versions.
multi-qa-mpnet-base-dot-v1 is the best performing sentence-transformers model for semantic search in English. It achieves an average performance of 62.18 across 14 datasets and requires 420 MB of storage. The model supports a maximum sequence length of 512 tokens and uses CLS pooling for dot-product scoring.written, reviewed, cited
