Arxiv:2506.14175 Base model:finetune:qwen/qwen3... Base model:qwen/qwen3-4b En Gram Qwen3 Region:us Reward Rewardmodel Rlhf Safetensors Sharded Tensorflow
Model Card on HF 🤗: https://huggingface.co/NiuTrans/GRAM-Qwen3-4B-RewardModel
GRAM Qwen3 4B RewardModel Benchmarks
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
What is the LLM Explorer Rank (Score)
GRAM Qwen3 4B RewardModel Parameters and Internals
| LLM Name | GRAM Qwen3 4B RewardModel |
| Repository 🤗 | https://huggingface.co/NiuTrans/GRAM-Qwen3-4B-RewardModel |
| Base Model(s) | |
| Model Size | 4b |
| Required VRAM | 8.1 GB |
| Updated | 2025-07-10 |
| Maintainer | NiuTrans |
| Model Type | qwen3 |
| Model Files | |
| Supported Languages | en |
| Model Architecture | Qwen3ForCausalLM |
| License | apache-2.0 |
| Context Length | 40960 |
| Model Max Length | 40960 |
| Transformers Version | 4.52.4 |
| Tokenizer Class | Qwen2Tokenizer |
| Padding Token | <|endoftext|> |
| Vocabulary Size | 151936 |
| Torch Data Type | bfloat16 |
| Errors | replace |
Best Alternatives to GRAM Qwen3 4B RewardModel
Best Alternatives |
Context / RAM |
Downloads |
Likes |
|---|---|---|---|
| Jan Nano 128K | 128K / 8.1 GB | 5031 | 197 |
| Polaris 4B Preview | 128K / 8.1 GB | 2407 | 113 |
| Jan Nano 128K | 128K / 8.1 GB | 215 | 2 |
| Jan Nano | 40K / 8.1 GB | 35986 | 442 |
| Qwen3 Embedding 4B | 40K / 8.1 GB | 92450 | 91 |
| Qwen3 Reranker 4B | 40K / 8.1 GB | 74828 | 74 |
| QVikhr 3 4B Instruction | 40K / 8.1 GB | 335 | 10 |
| Qwen3 4B | 40K / 8.1 GB | 18496 | 11 |
| Qwen3 4B FP8 | 40K / 5.2 GB | 6748 | 23 |
| Jan Nano | 40K / 8.1 GB | 549 | 5 |
Note: green Score (e.g. "73.2") means that the model is better than NiuTrans/GRAM-Qwen3-4B-RewardModel .
Rank the GRAM Qwen3 4B RewardModel Capabilities
🆘 Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! 🌟
What open-source LLMs or SLMs are you in search of? 49496 in total.
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241124