commit 106d0cb9884b88e2cbb2b8616cb1ba256f4e0bf2 Author: ModelHub XC Date: Wed Sep 16 22:34:29 2026 +0800 初始化项目，由ModelHub XC社区提供模型 Model: seanbailey518/Nanbeige4.1-3B-GGUF Source: Original Platform diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..4d1a820 --- /dev/null +++ b/.gitattributes @@ -0,0 +1,58 @@ +*.7z filter=lfs diff=lfs merge=lfs -text +*.arrow filter=lfs diff=lfs merge=lfs -text +*.bin filter=lfs diff=lfs merge=lfs -text +*.bz2 filter=lfs diff=lfs merge=lfs -text +*.ckpt filter=lfs diff=lfs merge=lfs -text +*.ftz filter=lfs diff=lfs merge=lfs -text +*.gz filter=lfs diff=lfs merge=lfs -text +*.h5 filter=lfs diff=lfs merge=lfs -text +*.joblib filter=lfs diff=lfs merge=lfs -text +*.lfs.* filter=lfs diff=lfs merge=lfs -text +*.mlmodel filter=lfs diff=lfs merge=lfs -text +*.model filter=lfs diff=lfs merge=lfs -text +*.msgpack filter=lfs diff=lfs merge=lfs -text +*.npy filter=lfs diff=lfs merge=lfs -text +*.npz filter=lfs diff=lfs merge=lfs -text +*.onnx filter=lfs diff=lfs merge=lfs -text +*.ot filter=lfs diff=lfs merge=lfs -text +*.parquet filter=lfs diff=lfs merge=lfs -text +*.pb filter=lfs diff=lfs merge=lfs -text +*.pickle filter=lfs diff=lfs merge=lfs -text +*.pkl filter=lfs diff=lfs merge=lfs -text +*.pt filter=lfs diff=lfs merge=lfs -text +*.pth filter=lfs diff=lfs merge=lfs -text +*.rar filter=lfs diff=lfs merge=lfs -text +*.safetensors filter=lfs diff=lfs merge=lfs -text +saved_model/**/* filter=lfs diff=lfs merge=lfs -text +*.tar.* filter=lfs diff=lfs merge=lfs -text +*.tar filter=lfs diff=lfs merge=lfs -text +*.tflite filter=lfs diff=lfs merge=lfs -text +*.tgz filter=lfs diff=lfs merge=lfs -text +*.wasm filter=lfs diff=lfs merge=lfs -text +*.xz filter=lfs diff=lfs merge=lfs -text +*.zip filter=lfs diff=lfs merge=lfs -text +*.zst filter=lfs diff=lfs merge=lfs -text +*tfevents* filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-BF16.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-Q4_0.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-Q4_1.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-Q5_0.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-Q5_1.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-TQ1_0.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-TQ2_0.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-Q2_K.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-IQ3_S.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-IQ3_M.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-Q3_K_S.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-Q3_K_M.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-Q3_K_L.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-IQ3_XS.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-IQ4_NL.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-IQ4_XS.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-Q4_K_S.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-Q4_K_M.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-Q5_K_S.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-Q5_K_M.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-Q6_K.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-Q8_0.gguf filter=lfs diff=lfs merge=lfs -text +Nanbeige4.1-3B-F16.gguf filter=lfs diff=lfs merge=lfs -text diff --git a/Nanbeige4.1-3B-BF16.gguf b/Nanbeige4.1-3B-BF16.gguf new file mode 100644 index 0000000..cca5636 --- /dev/null +++ b/Nanbeige4.1-3B-BF16.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0af03d1044c53502bd8ce99ba90136b2d153000fa72bdb100307d949a9d1f009 +size 7871577216 diff --git a/Nanbeige4.1-3B-F16.gguf b/Nanbeige4.1-3B-F16.gguf new file mode 100644 index 0000000..da6d305 --- /dev/null +++ b/Nanbeige4.1-3B-F16.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6213e20a73efa2b69c93524c25912d869fc73b691a98e485ca6466337a531964 +size 7871577216 diff --git a/Nanbeige4.1-3B-IQ3_M.gguf b/Nanbeige4.1-3B-IQ3_M.gguf new file mode 100644 index 0000000..e3b5db0 --- /dev/null +++ b/Nanbeige4.1-3B-IQ3_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c3260ebcef95c55b2ee35eda89c2f230db3166381a67ff1ba95d942aa3550822 +size 1908640896 diff --git a/Nanbeige4.1-3B-IQ3_S.gguf b/Nanbeige4.1-3B-IQ3_S.gguf new file mode 100644 index 0000000..4e91806 --- /dev/null +++ b/Nanbeige4.1-3B-IQ3_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b33af95b713b1a993e0bf5bfd76a5daacb5166c9e4790012be05c9ae47fadf5b +size 1866513536 diff --git a/Nanbeige4.1-3B-IQ3_XS.gguf b/Nanbeige4.1-3B-IQ3_XS.gguf new file mode 100644 index 0000000..1cd5d5f --- /dev/null +++ b/Nanbeige4.1-3B-IQ3_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6b145ec64a1b70a4e3a4d7633d803f4f4f9287259c2603ca465126ade887ea45 +size 1794260096 diff --git a/Nanbeige4.1-3B-IQ4_NL.gguf b/Nanbeige4.1-3B-IQ4_NL.gguf new file mode 100644 index 0000000..54352e5 --- /dev/null +++ b/Nanbeige4.1-3B-IQ4_NL.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d19e8675e8b8d71e84b6345bd9afe6f30b0ee679fac90cb9dae5dd34edd78c78 +size 2345545856 diff --git a/Nanbeige4.1-3B-IQ4_XS.gguf b/Nanbeige4.1-3B-IQ4_XS.gguf new file mode 100644 index 0000000..ed00b8d --- /dev/null +++ b/Nanbeige4.1-3B-IQ4_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4aa4ab0806b1fbab580ceb47c9d64e4708840e7c950e7661195005ba4e0c268f +size 2240585856 diff --git a/Nanbeige4.1-3B-Q2_K.gguf b/Nanbeige4.1-3B-Q2_K.gguf new file mode 100644 index 0000000..c5456d9 --- /dev/null +++ b/Nanbeige4.1-3B-Q2_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:31d6085bb0c8c4b83989412a6dc0339923f038aef516d03cab1e2cdce3008265 +size 1623103616 diff --git a/Nanbeige4.1-3B-Q3_K_L.gguf b/Nanbeige4.1-3B-Q3_K_L.gguf new file mode 100644 index 0000000..119029e --- /dev/null +++ b/Nanbeige4.1-3B-Q3_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0e1acb11de27cc56b25888e32a3a6fcc1783d9228470f903a36f90204a0d2f67 +size 2147499136 diff --git a/Nanbeige4.1-3B-Q3_K_M.gguf b/Nanbeige4.1-3B-Q3_K_M.gguf new file mode 100644 index 0000000..11c4d20 --- /dev/null +++ b/Nanbeige4.1-3B-Q3_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:30634bc25c78a011eb9835fbc420d795f335309ee47333e065ace571f66bc73a +size 2015607936 diff --git a/Nanbeige4.1-3B-Q3_K_S.gguf b/Nanbeige4.1-3B-Q3_K_S.gguf new file mode 100644 index 0000000..2947d82 --- /dev/null +++ b/Nanbeige4.1-3B-Q3_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d160ccd204f415006b65246ebfa3491b10d0e1c97b8779651de4eab2c57f90c0 +size 1860942976 diff --git a/Nanbeige4.1-3B-Q4_0.gguf b/Nanbeige4.1-3B-Q4_0.gguf new file mode 100644 index 0000000..4d0af10 --- /dev/null +++ b/Nanbeige4.1-3B-Q4_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:74212de616dc6bc7e10e01ea915ca5d28c81be51522b04f2aef339044280760b +size 2326868096 diff --git a/Nanbeige4.1-3B-Q4_1.gguf b/Nanbeige4.1-3B-Q4_1.gguf new file mode 100644 index 0000000..0d7d451 --- /dev/null +++ b/Nanbeige4.1-3B-Q4_1.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f43fe1d867222c196f11658b72d6a9e26afeb39530c9a28f15e02fe4d9509c49 +size 2546126976 diff --git a/Nanbeige4.1-3B-Q4_K_M.gguf b/Nanbeige4.1-3B-Q4_K_M.gguf new file mode 100644 index 0000000..1e13790 --- /dev/null +++ b/Nanbeige4.1-3B-Q4_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:acef47a794287be5b0f5e7387b42c660f02b4f79ea1f589bf55422eeed4969c2 +size 2443112576 diff --git a/Nanbeige4.1-3B-Q4_K_S.gguf b/Nanbeige4.1-3B-Q4_K_S.gguf new file mode 100644 index 0000000..79b6afb --- /dev/null +++ b/Nanbeige4.1-3B-Q4_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f9d0d96d972b5aa2a06b149d6e596407590a16ecec0c203f8610b45554a0ebc8 +size 2340958336 diff --git a/Nanbeige4.1-3B-Q5_0.gguf b/Nanbeige4.1-3B-Q5_0.gguf new file mode 100644 index 0000000..ffc4b01 --- /dev/null +++ b/Nanbeige4.1-3B-Q5_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:26227f8cc02cebfbd2b56c7d0579db4e66f57808e50977343a26fb024b9af5f2 +size 2765385856 diff --git a/Nanbeige4.1-3B-Q5_1.gguf b/Nanbeige4.1-3B-Q5_1.gguf new file mode 100644 index 0000000..b164ee2 --- /dev/null +++ b/Nanbeige4.1-3B-Q5_1.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:57fd3d53083c4ce3c9508fbe1bc37e77c9cc0060a3f557d9ee91a1af19e5ac96 +size 2984644736 diff --git a/Nanbeige4.1-3B-Q5_K_M.gguf b/Nanbeige4.1-3B-Q5_K_M.gguf new file mode 100644 index 0000000..448250f --- /dev/null +++ b/Nanbeige4.1-3B-Q5_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9e67b68d7af102548c5ef868c2c3725797798fc5505d49bae9bae90c827d20ab +size 2825269376 diff --git a/Nanbeige4.1-3B-Q5_K_S.gguf b/Nanbeige4.1-3B-Q5_K_S.gguf new file mode 100644 index 0000000..3e1f7ce --- /dev/null +++ b/Nanbeige4.1-3B-Q5_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5ef8cc6a10934a37e49ca138cdf39206afe483bed42f4bca37d72f8914c60276 +size 2765385856 diff --git a/Nanbeige4.1-3B-Q6_K.gguf b/Nanbeige4.1-3B-Q6_K.gguf new file mode 100644 index 0000000..3021741 --- /dev/null +++ b/Nanbeige4.1-3B-Q6_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fa37da6d9047e81affd096d8bb7f297e2fb408ead9938b058aa7cbdd9f440175 +size 3231310976 diff --git a/Nanbeige4.1-3B-Q8_0.gguf b/Nanbeige4.1-3B-Q8_0.gguf new file mode 100644 index 0000000..49bc027 --- /dev/null +++ b/Nanbeige4.1-3B-Q8_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5ab6d18d3c433b756ec35bcfaf08d2392cde5010b099c8e2f6ac9ae815b8f495 +size 4183948416 diff --git a/Nanbeige4.1-3B-TQ1_0.gguf b/Nanbeige4.1-3B-TQ1_0.gguf new file mode 100644 index 0000000..3a59c2f --- /dev/null +++ b/Nanbeige4.1-3B-TQ1_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dc8a92848b0c85bf0b16b5055c2db8d21d168ad190d45f4ccc5060bc1dd8bf79 +size 1243066496 diff --git a/Nanbeige4.1-3B-TQ2_0.gguf b/Nanbeige4.1-3B-TQ2_0.gguf new file mode 100644 index 0000000..feb61ad --- /dev/null +++ b/Nanbeige4.1-3B-TQ2_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8ea726f212afd0dfc7f538ddfedd9c6158910db1b50ed5364c1ec8ce0ea229ba +size 1387573376 diff --git a/README.md b/README.md new file mode 100644 index 0000000..a9d9527 --- /dev/null +++ b/README.md @@ -0,0 +1,268 @@ +--- +license: other +license_name: same-as-original +license_link: https://huggingface.co/Nanbeige/Nanbeige4.1-3B +base_model: Nanbeige/Nanbeige4.1-3B +tags: +- gguf +- quantized +- llama.cpp +--- + +# Nanbeige4.1-3B-GGUF + +This repository contains GGUF quantizations of [Nanbeige/Nanbeige4.1-3B](https://huggingface.co/Nanbeige/Nanbeige4.1-3B). + +## Quantizations + +| Filename | Size | Description | +|----------|------|-------------|| Nanbeige4.1-3B-TQ1_0.gguf | 1.16 GB | TQ1_0 | +| Nanbeige4.1-3B-TQ2_0.gguf | 1.29 GB | TQ2_0 | +| Nanbeige4.1-3B-Q2_K.gguf | 1.51 GB | Q2_K | +| Nanbeige4.1-3B-IQ3_XS.gguf | 1.67 GB | IQ3_XS | +| Nanbeige4.1-3B-Q3_K_S.gguf | 1.73 GB | Q3_K_S | +| Nanbeige4.1-3B-IQ3_S.gguf | 1.74 GB | IQ3_S | +| Nanbeige4.1-3B-IQ3_M.gguf | 1.78 GB | IQ3_M | +| Nanbeige4.1-3B-Q3_K_M.gguf | 1.88 GB | Q3_K_M | +| Nanbeige4.1-3B-Q3_K_L.gguf | 2 GB | Q3_K_L | +| Nanbeige4.1-3B-IQ4_XS.gguf | 2.09 GB | IQ4_XS | +| Nanbeige4.1-3B-Q4_0.gguf | 2.17 GB | Q4_0 | +| Nanbeige4.1-3B-Q4_K_S.gguf | 2.18 GB | Q4_K_S | +| Nanbeige4.1-3B-IQ4_NL.gguf | 2.18 GB | IQ4_NL | +| Nanbeige4.1-3B-Q4_K_M.gguf | 2.28 GB | Q4_K_M | +| Nanbeige4.1-3B-Q4_1.gguf | 2.37 GB | Q4_1 | +| Nanbeige4.1-3B-Q5_0.gguf | 2.58 GB | Q5_0 | +| Nanbeige4.1-3B-Q5_K_S.gguf | 2.58 GB | Q5_K_S | +| Nanbeige4.1-3B-Q5_K_M.gguf | 2.63 GB | Q5_K_M | +| Nanbeige4.1-3B-Q5_1.gguf | 2.78 GB | Q5_1 | +| Nanbeige4.1-3B-Q6_K.gguf | 3.01 GB | Q6_K | +| Nanbeige4.1-3B-Q8_0.gguf | 3.9 GB | Q8_0 | +| Nanbeige4.1-3B-F16.gguf | 7.33 GB | F16 | +| Nanbeige4.1-3B-BF16.gguf | 7.33 GB | BF16 | + +## Usage + +### With llama.cpp + +```bash +# Download a quantization +hf download seanbailey518/Nanbeige4.1-3B-GGUF --local-dir . + +# Run with llama.cpp +./llama-cli -m -p "Your prompt here" +``` + +### With llama-cpp-python + +```python +from llama_cpp import Llama + +llm = Llama(model_path="") +output = llm("Your prompt here", max_tokens=256) +print(output["choices"][0]["text"]) +``` + +## Quantization Details + +These quantizations were created using [llama.cpp](https://github.com/ggml-org/llama.cpp). + +- **Base Model:** [Nanbeige/Nanbeige4.1-3B](https://huggingface.co/Nanbeige/Nanbeige4.1-3B) +- **Quantization Tool:** llama.cpp convert_hf_to_gguf.py + llama-quantize + +## Original Model Card + +--- + +--- +license: apache-2.0 +language: +- en +- zh +library_name: transformers +pipeline_tag: text-generation +tags: +- llm +- nanbeige +base_model: +- Nanbeige/Nanbeige4-3B-Base +--- +
+ +Nanbeige Logo +
+ + + +# Introduction + +Nanbeige4.1-3B is built upon Nanbeige4-3B-Base and represents an enhanced iteration of our previous reasoning model, Nanbeige4-3B-Thinking-2511, achieved through further post-training optimization with supervised fine-tuning (SFT) and reinforcement learning (RL). As a highly competitive open-source model at a small parameter scale, Nanbeige4.1-3B illustrates that compact models can simultaneously achieve robust **reasoning**, **preference alignment**, and **effective agentic behaviors**. + +
+ + +
+ +Specifically, Nanbeige4.1-3B exhibits the following key strengths: + +* **Strong Reasoning:** Nanbeige4.1-3B is capable of solving complex, multi-step problems through sustained and coherent reasoning within a single forward pass, and reliably produces correct final answers on challenging tasks such as LiveCodeBench-Pro, IMO-Answer-Bench, and AIME 2026 I. +* **Robust Preference Alignment:** Nanbeige4.1-3B achieves solid alignment performance, outperforming not only same-scale models such as Qwen3-4B-2507 and Nanbeige4-3B-2511, but also substantially larger models including Qwen3-30B-A3B and Qwen3-32B on Arena-Hard-v2 and Multi-Challenge. +* **Agentic Capability:** Nanbeige4.1-3B is the first general small model to natively support deep-search tasks and reliably sustain complex problem solving involving more than 500 rounds of tool invocations. It fills a long-standing gap in the small-model ecosystem where models are typically optimized for either general reasoning or agentic scenarios, but rarely excel at both. + +> **Technical Report:** Coming Soon + + + + +# Performances + +We evaluate Nanbeige4.1-3B across a broad and diverse set of benchmarks covering **general reasoning**, and **deep-search capabilities**. + +### General Reasoning Tasks + +On general reasoning tasks including **code**, **math**, **science**, **alignment**, and **tool-use** benchmarks, Nanbeige4.1-3B not only significantly outperforms same-scale models such as **Qwen3-4B**, but also demonstrates overall superior performance compared to larger models including **Qwen3-30B-A3B-2507** and **Qwen3-32B**. + + +| Benchmark | Qwen3-4B-2507 | Qwen3-8B | Qwen3-14B | Qwen3-32B | Qwen3-30B-A3B-2507 | Nanbeige4-3B-2511 | **Nanbeige4.1-3B** | +| --------------------------- | ------------- | -------- | --------- | --------- | ------------------ | ----------------- | ------------------ | +| **Code** | | | | | | | | +| Live-Code-Bench-V6 | 57.4 | 49.4 | 55.9 | 55.7 | 66.0 | 46.0 | **76.9** | +| Live-Code-Bench-Pro-Easy | 40.2 | 41.2 | 33.0 | 42.3 | 60.8 | 40.2 | **81.4** | +| Live-Code-Bench-Pro-Mediium | 5.3 | 3.5 | 1.8 | 3.5 | 3.5 | 5.3 | **28.1** | +| **Math** | | | | | | | | +| AIME 2026 I | 81.46 | 70.42 | 76.46 | 75.83 | 87.30 | 84.1 | **87.40** | +| HMMT Nov | 68.33 | 48.33 | 56.67 | 57.08 | 71.25 | 66.67 | **77.92** | +| IMO-Answer-Bench | 48.00 | 36.56 | 41.81 | 43.94 | **54.34** | 38.25 | 53.38 | +| **Science** | | | | | | | | +| GPQA | 65.8 | 62.0 | 63.38 | 68.4 | 73.4 | 82.2 | **83.8** | +| HLE (Text-only) | 6.72 | 5.28 | 7.00 | 9.31 | 11.77 | 10.98 | **12.60** | +| **Alignment** | | | | | | | | +| Arena-Hard-v2 | 34.9 | 26.3 | 36.9 | 56.0 | 60.2 | 60.0 | **73.2** | +| Multi-Challenge | 41.14 | 36.30 | 36.97 | 38.72 | 49.40 | 41.20 | **52.21** | +| **Tool Use** | | | | | | | | +| BFCL-V4 | 44.87 | 42.20 | 45.14 | 47.90 | 48.6 | 53.8 | **56.50** | +| Tau2-Bench | 45.9 | 42.06 | 44.96 | 45.26 | 47.70 | 41.77 | **48.57** | + + + +### Deep Search Tasks + +As a general small model, Nanbeige4.1-3B achieves deep-search performance comparable to specialized agents under 10B parameters. +In contrast to existing small general models, which typically exhibit little to no deep-search capability, Nanbeige4.1-3B represents a substantial qualitative improvement over prior small general models. + +#### Deep Search and Agent Benchmarks +| Model | xBench-DeepSearch-2505 | xBench-DeepSearch-2510 | Browse-Comp | Browse-Comp-ZH | GAIA (Text-only) | HLE | SEAL-0 | +|------|-------------------|-------------------|-------------|----------------|------------------|-----|--------| +| **Search-Specialized Small Agents** |||||||| +| MiroThinker-v1.0-8B | 61 | – | 31.1 | 40.2 | 66.4 | 21.5 | 40.4 | +| AgentCPM-Explore-4B | 70 | – | 25.0 | 29.0 | 63.9 | 19.1 | 40.0 | +| **Large Foundation Models (with Tools)** |||||||| +| GLM-4.6-357B | 70 | – | 45.1 | 49.5 | 71.9 | 30.4 | – | +| Minimax-M2-230B | 72 | – | 44.0 | 48.5 | 75.7 | 31.8 | – | +| DeepSeek-V3.2-671B | 71 | – | 67.6 | 65.0 | 63.5 | 40.8 | 38.5 | +| **Small Foundation Models (with Tools)** |||||||| +| Qwen3-4B-2507 | 34 | 5 | 1.57 | 7.92 | 28.33 | 11.13 | 15.74 | +| Qwen3-8B | 31 | 2 | 0.79 | 5.15 | 19.53 | 10.24 | 6.34 | +| Qwen3-14B | 34 | 9 | 2.36 | 7.11 | 30.23 | 10.17 | 12.64 | +| Qwen3-32B | 39 | 8 | 3.15 | 7.34 | 30.17 | 9.26 | 8.15 | +| Qwen3-30B-A3B-2507 | 25 | 10| 1.57 | 4.12 | 31.63 | 14.81 | 9.24 | +| **Ours (with Tools)** |||||||| +| Nanbeige4-3B-2511 | 33 | 11 | 0.79 | 3.09 | 19.42 | 13.89 | 12.61 | +| **Nanbeige4.1-3B** | **75** | **39** | **19.12** | **31.83** | **69.90** | **22.29** | **41.44** | + + +## Quickstart + +For inference hyperparameters, we recommend the following settings: +* Temperature: 0.6 +* Top-p: 0.95 +* Repeat penalty: 1.0 +* Max New Tokens: 131072 + +For the chat scenario: +``` +from transformers import AutoModelForCausalLM, AutoTokenizer +tokenizer = AutoTokenizer.from_pretrained( + 'Nanbeige/Nanbeige4.1-3B', + use_fast=False, + trust_remote_code=True +) +model = AutoModelForCausalLM.from_pretrained( + 'Nanbeige/Nanbeige4.1-3B', + torch_dtype='auto', + device_map='auto', + trust_remote_code=True +) +messages = [ + {'role': 'user', 'content': 'Which number is bigger, 9.11 or 9.8?'} +] +prompt = tokenizer.apply_chat_template( + messages, + add_generation_prompt=True, + tokenize=False +) +input_ids = tokenizer(prompt, add_special_tokens=False, return_tensors='pt').input_ids +output_ids = model.generate(input_ids.to('cuda'), eos_token_id=166101) +resp = tokenizer.decode(output_ids[0][len(input_ids[0]):], skip_special_tokens=True) +print(resp) +``` + +For the tool use scenario: +``` +from transformers import AutoModelForCausalLM, AutoTokenizer +tokenizer = AutoTokenizer.from_pretrained( + 'Nanbeige/Nanbeige4.1-3B', + use_fast=False, + trust_remote_code=True +) +model = AutoModelForCausalLM.from_pretrained( + 'Nanbeige/Nanbeige4.1-3B', + torch_dtype='auto', + device_map='auto', + trust_remote_code=True +) +messages = [ + {'role': 'user', 'content': 'Help me check the weather in Beijing now'} +] +tools = [{'type': 'function', + 'function': {'name': 'SearchWeather', + 'description': 'Find out the current weather in a place on a certain day.', + 'parameters': {'type': 'dict', + 'properties': {'location': {'type': 'string', + 'description': 'A city in China.'}, + 'required': ['location']}}}}] +prompt = tokenizer.apply_chat_template( + messages, + tools, + add_generation_prompt=True, + tokenize=False +) +input_ids = tokenizer(prompt, add_special_tokens=False, return_tensors='pt').input_ids +output_ids = model.generate(input_ids.to('cuda'), eos_token_id=166101) +resp = tokenizer.decode(output_ids[0][len(input_ids[0]):], skip_special_tokens=True) +print(resp) +``` + +For the deep-search scenario: + +* Inference Framework: [**miroflow-framework**](https://github.com/MiroMindAI/MiroThinker)! +* Switch tokenizer configuration to **tokenizer_config_search.json** +* Tools Configuration: + +| Server | Description | Tools Provided | +|-----------------------------|-----------------------------------------------------------------------------|-------------------------------------------------------------------------------| +| tool-python | Execution environment and file management ([E2B sandbox](https://e2b.dev/)) | create_sandbox, run_command, run_python_code, upload_file_from_local_to_sandbox, download_file_from_sandbox_to_local, download_file_from_internet_to_sandbox | +| search_and_scrape_webpage | Google search via [Serper API](https://google.serper.dev) | google_search | +| jina_scrape_llm_summary | Web scraping with LLM-based information extraction with [Jina](https://r.jina.ai) | scrape_and_extract_info | + +* Summary model: Qwen3-14B-thinking +* Temperature: 1.0 +* Note, access to HuggingFace has been explicitly disabled in these tools. + +# Limitations + +While we place great emphasis on the safety of the model during the training process, striving to ensure that its outputs align with ethical and legal requirements, it may not completely avoid generating unexpected outputs due to the model's size and probabilistic nature. These outputs may include harmful content such as bias or discrimination. Please don't propagate such content. We do not assume any responsibility for the consequences resulting from the dissemination of inappropriate information. +
+ +# Contact +If you have any questions, please raise an issue or contact us at nanbeige@kanzhun.com. +