gte-reranker-modernbert-base

Alibaba-NLP Apache 2.0 out 21 months ago

150M parameters, a query and passages in, relevance scores out. 1 provider serves it.

Verified 2026-10-25 Rank #41 in Embeddings · score 91 method v1 Sources 3 Downloads 30d 1.7M · #71 of 619 tracked

Changed this week: 2 new quantized builds on the Hub (9 in all).2026-10-24

Can I run it?0.1GB at INT4Estimatefits a 16 GB cardmethod

Can I use it?Apache 2.0Permissive; patent grantlicence file

Do people use it?1.7Mdownloads, 30 days#71 of 619 we track · top 11%+4 likes, downloads −16% this week110 likesHugging Face

Where to run it

Live hosted endpoints, cheapest first, USD per million tokens. Ties break on output price, then provider name.

#	Cheapest live row	In /1M	Out /1M	Context	Quant	Seen
1	HF Inference	—	—	—	—	Oct 25

All 1 provider →

Where it's used

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 1 API provider; 9 quantized builds for local use.

As a service1HF Inference
On own hardware9quantized builds
Built on2424 fine-tunes · 0 adapters · 0 merges
In apps0Hugging Face Spaces

Will it fit?

Weights only, an estimate: leave room for the KV cache. Method

Fits a 16 GB card at INT4

About 0.1 GB at INT4, 0.2 GB at FP8, 0.4 GB at BF16: weights only, so leave room for the KV cache.

Estimate params × bytes per param × 1.2 · how we estimate

What people built on it

33 derivatives on the Hub. Which build to run, by runtime: the most downloaded in each class.

33 derivatives on the Hub, most of them fine-tunes

From the Hugging Face model tree; quantized builds in teal, fine-tunes, merges and adapters in greys.

GGUF jolleyboy/gte-reranker-modernbert-base-GGUF for llama.cpp, Ollama, LM Studio 557 downloads

All variants →

Reference

Everything else, folded: the full tables, every source.

Use-case evidence

4 use cases

What it is good at 4 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.

Facts

15 fields

Listed in Embeddings

Facts

gte-reranker-modernbert-base has 150M parameters. (safetensors metadata, observed —)
License: Apache 2.0 (friction: none). Permissive; patent grant. (license file, observed —)
Created 2025-01-20, last modified 2025-07-04. (HF API, observed —)

Pipeline tag	text-ranking
Library	transformers
Architectures	ModernBertForSequenceClassification
Model type	modernbert
Params by dtype	F32: 149,605,633
Size band	S
MoE	no
Papers	arXiv:2308.03281
License	Apache 2.0 (apache-2.0)
Hub snapshot	2026-10-25

Hardware fit

3 estimates

Hardware fit ESTIMATE

Weights-only memory for gte-reranker-modernbert-base: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.

BF16 (2 bytes/param)	0.4 GB · Edge (<4 GB)
FP8 (1 byte/param)	0.2 GB · Edge (<4 GB)
INT4 (0.5 bytes/param)	0.1 GB · Edge (<4 GB)

Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra.

All variants

33 variants

Lineage

Fine-tuned from answerdotai/ModernBERT-base Family: Alibaba-NLP/gte-reranker 33 derivatives on the Hub

Base model: answerdotai/ModernBERT-base

finetune: 24quantized: 9

Top variants (33 shown, ranked by downloads)

Variant	Relation	Quant	Params	Band	License	Author	Likes	Downloads
KRLabsOrg/verbatim-rag-modern-bert-v2	finetune	—	150M	S	apache-2.0	KRLabsOrg	13	10.5k
redis/langcache-reranker-v1-bce	finetune	—	150M	S	apache-2.0	redis	0	3.8k
hotchpotch/open-provence-reranker-v1-gte-modernbert-base	finetune	—	150M	S	mit	hotchpotch	2	2.5k
redis/langcache-reranker-v2-mnrl	finetune	—	150M	S	apache-2.0	redis	3	1.4k
amanwithaplan/arcade-reranker	finetune	—	150M	S	—	amanwithaplan	0	934
redis/langcache-reranker-v2-bce	finetune	—	150M	S	apache-2.0	redis	0	831
redis/langcache-reranker-v1-mnrl	finetune	—	150M	S	apache-2.0	redis	0	755
jolleyboy/gte-reranker-modernbert-base-GGUF	quantized	GGUF	—	—	apache-2.0	jolleyboy	0	557
kasraahmadi/job-eval-modernbert	finetune	—	156M	S	apache-2.0	kasraahmadi	1	412
yjoonjang/preranker-v1	finetune	—	150M	S	apache-2.0	yjoonjang	4	228
keisuke-miyako/gte-reranker-modernbert-base-gguf-q8_0	quantized	gguf	—	—	—	keisuke-miyako	0	393
KRLabsOrg/acl-verbatim-modernbert	finetune	—	150M	S	apache-2.0	KRLabsOrg	3	186
afafos/trade-news-dedup-gte-reranker-modernbert-base	finetune	—	150M	S	apache-2.0	afafos	0	316
IDQO/arcade-reranker	finetune	—	150M	S	—	IDQO	0	269
amanwithaplan/arcade-reranker-1500	finetune	—	150M	S	—	amanwithaplan	0	266
keisuke-miyako/gte-reranker-modernbert-base-gguf-q4_k_m	quantized	gguf	—	—	—	keisuke-miyako	0	182
NYSgpt/44b-reranker	finetune	—	150M	S	apache-2.0	NYSgpt	2	60
keisuke-miyako/gte-reranker-modernbert-base-gguf-f16	quantized	gguf	—	—	—	keisuke-miyako	0	124
SMARTICT/gte-reranker-modernbert-base-pubmed-tr-v1	finetune	—	150M	S	apache-2.0	SMARTICT	0	95
PSanni/gte-modernbert-reranker-custom-max1	finetune	—	150M	S	—	PSanni	0	93
jayshah5696/er2-ce-gte-reranker-ft	finetune	—	150M	S	—	jayshah5696	0	88
Mithilss/gte-reranker-modernbert-base-finetune	finetune	—	150M	S	apache-2.0	Mithilss	0	85
thi3rrr/gte-reranker-modernbert-base-GGUF	quantized	GGUF	—	—	apache-2.0	thi3rrr	0	80
PSanni/gte-modernbert-reranker-custom	finetune	—	150M	S	—	PSanni	0	77
rahulydc/verbatim-rag-modern-bert-v2	finetune	—	150M	S	apache-2.0	rahulydc	0	74
keisuke-miyako/gte-reranker-modernbert-base-onnx-int8	quantized	int8	—	—	—	keisuke-miyako	0	65
nehajiya8/esci-gte-reranker	finetune	—	150M	S	—	nehajiya8	0	64
the-peoples-solutions/gte-reranker-modernbert-base-GGUF	quantized	GGUF	—	—	apache-2.0	the-peoples-solutions	0	64
Sgupta02/code-search-reranker-finetuned	finetune	—	150M	S	apache-2.0	Sgupta02	0	58
keisuke-miyako/gte-reranker-modernbert-base-onnx-f16	quantized	—	—	—	—	keisuke-miyako	0	55
akzaidan/People2JobRanker	finetune	—	—	—	mit	akzaidan	1	0
freeideas/UsefulLocalModels	quantized	GGUF	—	—	apache-2.0	freeideas	0	47
TheMattFace/gte-reranker-naga-toolpicker	finetune	—	150M	S	apache-2.0	TheMattFace	0	10

Repository changelog

1 change

changed in the last 7 daysOther: 2026-10-24all changes

All providers

1 provider · 1 endpoint

Where to run 1 providers · 1 endpoints

Hosted endpoints observed for gte-reranker-modernbert-base, sorted by input price (USD per million tokens). Each row names its observer and date; prices change daily, so check the source before quoting.

Provider	Model ref	Context	Quant	$/M in	$/M out	tok/s	TTFT	Tools	Status	Observed
HF Inference hf	Alibaba-NLP/gte-reranker-modernbert-base	—	—	—	—	—	—	—	live	2026-10-25

Articles and coverage

4 pieces

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.

Analysis4

Alibaba GTE Reranker ModernBERT-base — PublicAI Index

publicai.io100 words

GTE-reranker-ModernBERT — long-context cross-encoder reranker in your browser · Web AI Showcase

webai.showabout the gte reranker family839 words

Alibaba's GTE-reranker-ModernBERT is a cross-encoder reranker built on the ModernBERT backbone that supports an 8192-token context. The 768 hidden, 22 layers, 12 heads architecture uses q8 (8-bit) ONNX quantization and carries an Apache-2.0 license. It runs entirely on-device via WebAssembly without needing WebGPU.written, reviewed, cited

Inference runs in a Web Worker to ensure ranking operations never jank the main page thread.
The tokenizer pads and truncates inputs to fit the 8192-token limit during the encoding step.

gte-reranker-modernbert-base | PromptLayer Models

promptlayer.com334 words

Alibaba's Tongyi Lab released gte-reranker-modernbert-base, a 149M parameter text reranker built on modernBERT that supports sequences up to 8192 tokens. The model uses Flash Attention 2 for efficiency and achieves a LoCo score of 90.68 and a BEIR score of 56.19. It outputs calibrated relevance scores for query-document pairs to refine search results.written, reviewed, cited

8192长文本极限优化！gte-reranker-modernbert-base本地部署与推理全流程（含ONNX量化加速）_皮强策Darcy-火山引擎 ADG 社区

adg.csdn.net1,026 words

The gte-reranker-modernbert-base model enables local deployment for long-text reranking tasks with a maximum sequence length of 8192 tokens. This 149M parameter model uses ModernBERT architecture with ALiBi position encoding to achieve industrial-grade performance while supporting ONNX quantization for faster inference. The optimized attention calculation reduces memory complexity from O(n²) to O(n√n) to handle extended contexts efficiently.written, reviewed, cited

ONNX INT8 quantization achieves 8.7ms inference per pair on RTX 4090 with 310MB memory usage.
Flash Attention acceleration requires GPU compute capability of 8.0 or higher.

Papers

1 paper

Papers 1

arXiv links found in the model card and its tags. "Peer-reviewed" is shown only with evidence: a published venue in OpenAlex (measured) or the authors' own journal reference on arXiv (reported).

Towards General Text Embeddings with Multi-stage Contrastive Learning cited in card

Zehan Li, Xin Zhang, Yanzhao Zhang +3 · 2023-08-07 · PDF · HF paper page (3 upvotes)

Similar models

3 models

Similar models

jina-reranker-v2-base-multilingual · 278M
same size band (S), same task (text-ranking)
nomic-embed-text-v1.5 · 137M
same size band (S), permissive license
bart-base · 139M
same size band (S), permissive license

Sources

3 sources

Sources

Hub snapshot: 2026-10-25 · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:45 UTC · Summary: 2026-10-14 14:07 UTC · Last verified 2026-10-25

Follow gte-reranker-modernbert-base

Watch this model

/feeds/models/Alibaba-NLP--gte-reranker-modernbert-base.xml(RSS 2.0) · every change to gte-reranker-modernbert-base
/feeds/families/Alibaba-NLP--gte-reranker.xml(RSS 2.0) · every change across the gte-reranker family

Email alerts are not open yet; the feeds above are the same data. See feeds for the item schema.