Search public pages, research tools, and SAVRN solutions.
ModernBERT-large is an open-weight model for fill mask from Answer.AI, released under Apache License 2.0. It has 396M parameters and a 8,192-token context. At 16-bit it needs about 1 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index. It draws 750.2k downloads a month.
ModernBERT is a modernized bidirectional encoder-only Transformer model (BERT-style) pre-trained on 2 trillion tokens of English and code data with a native context length of up to 8,192 tokens.
What it takes to serve ModernBERT-large (396M parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.
| Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits |
|---|---|---|---|---|---|
| 16-bit | 0.8 GB | 1.0 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 0.4 GB | 0.5 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 0.2 GB | 0.2 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Oct 9, 2026.
ModernBERT-large on every accelerator the SAVRN Index prices, at every precision
Nothing here generates text. Answer.AI built ModernBERT-large as an encoder, trained on 2 trillion tokens of English and code, and our reading is retrieval and classification over long documents, up to its 8,192-token context. Memory is the smallest line item in the rack: 1.0 GB at 16-bit, 0.2 GB at 4-bit, against a cheapest slot of one MI300X with 192 GB at $1.85 an hour, so it rides alongside a generation model rather than earning its own card.
The license is Apache 2.0: commercial use, modification and redistribution are allowed, notices stay, significant changes get stated, and contributors grant patents. Access is open. Check the disk footprint first, 16 files and 7.86 GB stored in float32, several times what loads into memory at 16-bit. It ships in safetensors, onnx and pytorch. The design paper is arXiv:2412.13663; read it for the local-global alternating attention behind the 8,192-token window.
By Answer.AI, published under apache-2.0, revision 45bb4654a4d5.
ModernBERT is a modernized bidirectional encoder-only Transformer model (BERT-style) pre-trained on 2 trillion tokens of English and code data with a native context length of up to 8,192 tokens. ModernBERT leverages recent architectural improvements such as:
ModernBERT’s native long context length makes it ideal for tasks that require processing long documents, such as retrieval, classification, and semantic search within large corpora. The model was trained on a large corpus of text and code, making it suitable for a wide range of downstream tasks, including code retrieval and hybrid (text + code) semantic search.
It is available in the following sizes:
16 files, 7.9 GB in total. The weights are 10 files totalling 7.9 GB in bin, onnx, safetensors.
| File | Type | Size | SHA-256 |
|---|---|---|---|
| model.safetensors | Weights | 1.6 GB | 44510fec5d3a |
| onnx/model.onnx | Weights | 1.6 GB | 6bab50009674 |
| onnx/model_bnb4.onnx | Weights | 400.9 MB | 8b179381ce3c |
| onnx/model_fp16.onnx | Weights | 792.2 MB | 29cef3547cd4 |
| onnx/model_int8.onnx | Weights | 398.3 MB | 84baefd65fac |
| onnx/model_q4.onnx | Weights | 422.4 MB | b480f884c31d |
| onnx/model_q4f16.onnx | Weights | 297.5 MB | 0a12b6229a1f |
| onnx/model_quantized.onnx | Weights | 398.3 MB | 84baefd65fac |
| onnx/model_uint8.onnx | Weights | 398.3 MB | 9020c46e799c |
| pytorch_model.bin | Weights | 1.6 GB | fabb014a0b62 |
| config.json | Configuration | 1.2 KB | — |
| special_tokens_map.json | Configuration | 694 B | — |
| README.md | Documentation | 8.4 KB | — |
| .gitattributes | Repository | 1.5 KB | — |
| tokenizer.json | Tokenizer | 2.1 MB | — |
| tokenizer_config.json | Tokenizer | 20.8 KB | — |
Released by Answer.AI through its official repository on Hugging Face. Read the license.
| Precision | Weights in memory |
|---|---|
| As published | 7.9 GB |
| 16-bit | 0.8 GB |
| 8-bit | 0.4 GB |
| 4-bit | 0.2 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
About 1 GB at 16-bit and 0.2 GB at 4-bit: the weights (396M parameters) plus a working margin. A long context needs more.
At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.
Yes. ModernBERT-large is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.
8,192 tokens, from the maximum position embeddings in its published configuration.
This model is a fine-tuned version of ModernBERT optimized for logical reasoning, deductive analysis, and structure-based token prediction. It relies on the ModernBertForMaskedLM architecture, making it highly effective at filling in missing contextual logic tokens (fill-mask). - As an encoder-based Masked Language Model, it is not designed for long-form generative text (like ChatGPT). It excels at predicting masked tokens inside structured prompts.
Pretrained model on English language using a masked language modeling (MLM) objective. It was introduced in this paper and first released in this repository. This model is case-sensitive: it makes a difference between english and English. Disclaimer: The team releasing RoBERTa did not write a model card for this model so this model card has been written by the Hugging Face team. RoBERTa is a transformers model pretrained on a large corpus of English data in a self-supervised fashion. This means it was pretrained on the raw texts only, with no humans labelling them in any way (which is why it can use lots of publicly available data) with an automatic process to generate inputs and labels…
Pretrained model on English language using a masked language modeling (MLM) objective. It was introduced in this paper and first released in this repository. This model is uncased: it does not make a difference between english and English. Disclaimer: The team releasing BERT did not write a model card for this model so this model card has been written by the Hugging Face team. BERT is a transformers model pretrained on a large corpus of English data in a self-supervised fashion. This means it was pretrained on the raw texts only, with no humans labelling them in any way (which is why it can use lots of publicly available data) with an automatic process to generate inputs and labels from…
This is a Japanese DeBERTa V2 large model pre-trained on Japanese Wikipedia, the Japanese portion of CC-100, and the Japanese portion of OSCAR. This model is trained with character-level tokenization and whole word masking. You can use this model for masked language modeling as follows: You can also fine-tune this model on downstream tasks. There is no need to tokenize texts in advance, and you can give raw texts to the tokenizer. The texts are tokenized into character-level tokens by sentencepiece. We used the following corpora for pre-training: - Japanese Wikipedia (as of 20221020, 3.2GB, 27M sentences, 1.3M documents) - Japanese portion of CC-100 (85GB, 619M sentences, 66M documents)…
This model is a fine-tuned version of jhu-clsp/mmBERT-base on an unknown dataset. It achieves the following results on the evaluation set: The following hyperparameters were used during training: - learningrate: 3e-05 - trainbatchsize: 8 - evalbatchsize: 4 - distributedtype: multi-GPU - numdevices: 2 - gradientaccumulationsteps: 2 - totaltrainbatchsize: 32 - totalevalbatchsize: 8 - lrschedulertype: cosine - lrschedulerwarmupsteps: 0.1 - numepochs: 4 - mixedprecisiontraining: Native AMP - Transformers 5.0.0 - Pytorch 2.10.0+cu128 - Datasets 5.0.0 - Tokenizers 0.22.2
XLM-RoBERTa model pre-trained on 2.5TB of filtered CommonCrawl data containing 100 languages. It was introduced in the paper Unsupervised Cross-lingual Representation Learning at Scale by Conneau et al. and first released in this repository. Disclaimer: The team releasing XLM-RoBERTa did not write a model card for this model so this model card has been written by the Hugging Face team. XLM-RoBERTa is a multilingual version of RoBERTa. It is pre-trained on 2.5TB of filtered CommonCrawl data containing 100 languages. RoBERTa is a transformers model pretrained on a large corpus in a self-supervised fashion. This means it was pretrained on the raw texts only, with no humans labelling them in…