---
license: apache-2.0
base_model: Qwen/Qwen3-8B
tags:
  - terminal
  - sft
  - marin
  - nemotron
  - qwen3
datasets:
  - nvidia/Nemotron-Terminal-Corpus
model-index:
  - name: nemotron-terminal-8b-repro
    results:
      - task:
          type: text-generation
          name: Terminal-Bench 2.0
        dataset:
          name: Terminal-Bench 2.0
          type: custom
        metrics:
          - type: accuracy
            value: 15.9
            name: TB2 Accuracy (%)
      - task:
          type: text-generation
          name: Terminal-Bench Lite
        dataset:
          name: OpenThoughts-TBLite
          type: open-thoughts/OpenThoughts-TBLite
        metrics:
          - type: accuracy
            value: 29.0
            name: TBLite Accuracy (%)
---

# Nemotron-Terminal-8B Reproduction (Marin)

A reproduction of [NVIDIA's Nemotron-Terminal-8B](https://huggingface.co/nvidia/Nemotron-Terminal-8B) by the [Marin community](https://github.com/marin-community/marin), SFT'd on the full [Nemotron-Terminal-Corpus](https://huggingface.co/datasets/nvidia/Nemotron-Terminal-Corpus) (366K examples).

## Results

| Benchmark | Nemotron-Terminal-8B (released) | This model (Marin SFT) | Diff |
|---|---|---|---|
| **Terminal-Bench 2.0** | 13.0% ± 2.2 | **15.9%** (14/88) | **+2.9pp** |
| **Terminal-Bench Lite** | 23.0% (23/100) | **29.0%** (29/100) | **+6.0pp** |

Both benchmarks show our reproduction outperforming the released weights, despite training on 366K examples vs the paper's 490K.

## Training Details

- **Base model**: [Qwen/Qwen3-8B](https://huggingface.co/Qwen/Qwen3-8B)
- **Training data**: [nvidia/Nemotron-Terminal-Corpus](https://huggingface.co/datasets/nvidia/Nemotron-Terminal-Corpus) — 366K examples (skill_based_easy, skill_based_medium, skill_based_mixed, diverse_complex subsets)
- **Steps**: 5,721 (2 epochs)
- **Learning rate**: 2e-5 (cosine schedule)
- **Batch size**: 128
- **Sequence length**: 32,768
- **Final loss**: ~0.35
- **Hardware**: TPU v5p-32
- **Framework**: [Marin](https://github.com/marin-community/marin) (Levanter-based)
- **W&B run**: [exp3490b](https://wandb.ai/marin-community/marin/runs/exp3490b_sft_nemotron_terminal_corpus_full_qwen3_8b_32768tokens_-3da6c1)

## Evaluation Setup

- **Inference**: vLLM-TPU (native) on v5p-8
- **Agent**: Terminus-2 (temperature=0.6)
- **Environments**: Daytona

## Links

- **Tracking issue**: [marin-community/marin#3490](https://github.com/marin-community/marin/issues/3490)
- **Training code**: [experiments/exp3490b](https://github.com/marin-community/marin/blob/kevin/harbor/experiments/exp3490b_sft_nemotron_terminal_corpus_full_qwen3_8b_32768tokens.py)
- **Eval code**: [experiments/exp3490b_eval](https://github.com/marin-community/marin/blob/kevin/harbor/experiments/exp3490b_eval_marin_sft_nemotron.py)