allenai / tmax-9b

huggingface.co
Total runs: 4.5K
24-hour runs: -133
7-day runs: -512
30-day runs: -22.2K
Model's Last Updated: June 23 2026

Introduction of tmax-9b

Model Details of tmax-9b

image

💻 Code · 🤗 Models & Data · 📜 Paper · 📓 Blog

For full information, go check out the Tmax paper here .

TMax 9B

TMax 9B is a model trained using DPPO on top of Qwen 3.5 9B for use as a terminal-agent. It achieves roughly 27% on Terminal Bench 2.0 after 200 steps of RL training, improving by ~6 points over Qwen 3.5 9B with our harness.

image

This model is part of a collection of terminal agents in various sizes.

Additionally, we provide model checkpoints as branches of the repository. The main model checkpoint is step 200 as this performed best on TBLite.

We also provide all model rollouts and logprobs over the course of training here ! This allows you to inspect what went on during our runs. See the folder for more details (I recommend pointing an agent at it to make plots, etc).

Evaluation Results
Model TB Lite TB 2.1 TB 2.0 (daytona)
Qwen 3.5 2B 5.71 +/- 1.6 1.9 +/- 1.4 2.3 +/- 1.0
Tmax 2B 11.8 +/- 1.4 4.2 +/- 1.2 2.9 +/- 0.6
Qwen 3.5 4B 31.8 +/- 3.8 ? 16.6 +/- 1.7
Tmax 4B 42.6 +/- 1.5 19.9 +/- 1.1 18.9 +/- 1.9
Qwen 3.5 9B 41.9 +/- 2.7 16.1 +/- 3.7 21.1 +/- 2.6
Tmax 9B (this model!) 57.2 +/- 2.5 28.8 +/- 3.7 27.2 +/- 1.5
Qwen 3.6 27B 70.8 +/- 2.1 40.5 +/- 2.4 39.6 +/- 2.1
Tmax 27B 68.6 +/- 4.7 44.9 +/- 1.8 42.7 +/- 0.7

For details on evaluation methodology please check our paper. In general, we used a podman (docker) backend with default timeouts and custom harness similar to mini-swe-agent. For the 'daytona' runs, we used the daytona backend. For Lite/2.1, we show mean and standard error over 3 runs. For daytona, we show it over 5 runs.

Model Details
Model Description
  • Developed by: Ai2
  • Language(s) (NLP): English
  • License: Apache 2.0
  • Finetuned from model [optional]: Qwen 3.5 9B
  • Dataset: TMax-15k
Use

To use this model, we recommend serving with vllm (or your inference framework of choice) with:

uvx vllm==0.19.1 serve allenai/tmax-9b \
  --served-model-name tmax-9b \
  --enable-auto-tool-choice \
  --tool-call-parser qwen3_xml \
  --port 8008 \
  --max-model-len 65536 \
  --tensor-parallel-size 8 \
  --language_model_only

Make sure to set language_model_only as we removed the vision head during training.

For more details on evaluation, please see our codebase .

Hyperparameters

This model was trained using DPPO with the following hyperparameters:

  • base model : hamishivi/Qwen3.5-9B
  • Dataset : tmax 15K
  • Max prompt tokens : 2048
  • Max per-turn tokens : 16384
  • Max overall tokens : 65536
  • Pack length : 67584
  • Per-device train batch size : 1
  • Unique prompts per rollout : 8
  • Samples per prompt rollout : 32
  • Async steps : 4
  • Max steps : 64
  • Learning rate : 1e-6
  • LR scheduler : constant
  • Total training steps : 500 steps (this checkpoint is from 200 steps of training, which performed best on TBLite)
  • Sampling Temperature : 1.0
  • KL Beta : 0.0
  • Loss fn : DPPO
  • Divergence : binary TV
  • TV threshold : 0.1
  • Advantage normalization : centered (no division by stdev)
  • FP32 LM head : true

For more details on training, please see our codebase .

License

This model is licensed under Apache 2.0. It is intended for research and educational use in accordance with Ai2's Responsible Use Guidelines .

Citation

If you use our model or data, please cite our paper:

@misc{ivison2026tmax,
  title={{TMAX}: A Simple Recipe for Terminal Agents},
  author={Ivison, Hamish and Yin, Junjie Oscar and Shao, Rulin and Xiao, Teng and Lambert, Nathan and Hajishirzi, Hannaneh},
  year={2026},
}

Runs of allenai tmax-9b on huggingface.co

4.5K
Total runs
-133
24-hour runs
-512
3-day runs
-512
7-day runs
-22.2K
30-day runs

More Information About tmax-9b huggingface.co Model

More tmax-9b license Visit here:

https://choosealicense.com/licenses/apache-2.0

tmax-9b huggingface.co

tmax-9b huggingface.co is an AI model on huggingface.co that provides tmax-9b's model effect (), which can be used instantly with this allenai tmax-9b model. huggingface.co supports a free trial of the tmax-9b model, and also provides paid use of the tmax-9b. Support call tmax-9b model through api, including Node.js, Python, http.

allenai tmax-9b online free

tmax-9b huggingface.co is an online trial and call api platform, which integrates tmax-9b's modeling effects, including api services, and provides a free online trial of tmax-9b, you can try tmax-9b online for free by clicking the link below.

allenai tmax-9b online free url in huggingface.co:

https://huggingface.co/allenai/tmax-9b

tmax-9b install

tmax-9b is an open source model from GitHub that offers a free installation service, and any user can find tmax-9b on GitHub to install. At the same time, huggingface.co provides the effect of tmax-9b install, users can directly use tmax-9b installed effect in huggingface.co for debugging and trial. It also supports api for free installation.

tmax-9b install url in huggingface.co:

https://huggingface.co/allenai/tmax-9b

Url of tmax-9b

tmax-9b huggingface.co Url

Provider of tmax-9b huggingface.co

allenai
ORGANIZATIONS

Other API from allenai

huggingface.co

Total runs: 1.1M
Run Growth: 355.2K
Growth Rate: 33.34%
Updated:December 04 2024
huggingface.co

Total runs: 146.4K
Run Growth: 19.3K
Growth Rate: 13.21%
Updated:July 28 2025
huggingface.co

Total runs: 136.2K
Run Growth: -2.6K
Growth Rate: -1.93%
Updated:January 23 2026
huggingface.co

Total runs: 114.5K
Run Growth: -141.3K
Growth Rate: -123.42%
Updated:January 23 2026
huggingface.co

Total runs: 100.0K
Run Growth: 37.8K
Growth Rate: 37.79%
Updated:January 23 2026
huggingface.co

Total runs: 55.7K
Run Growth: 53.5K
Growth Rate: 95.99%
Updated:October 10 2025
huggingface.co

Total runs: 38.2K
Run Growth: 10.4K
Growth Rate: 27.18%
Updated:August 15 2024
huggingface.co

Total runs: 21.6K
Run Growth: 4.8K
Growth Rate: 22.37%
Updated:October 18 2023
huggingface.co

Total runs: 11.5K
Run Growth: 7.5K
Growth Rate: 65.13%
Updated:May 23 2026
huggingface.co

Total runs: 7.2K
Run Growth: 1.3K
Growth Rate: 18.08%
Updated:April 11 2026
huggingface.co

Total runs: 5.1K
Run Growth: 1.2K
Growth Rate: 23.17%
Updated:December 04 2024
huggingface.co

Total runs: 4.4K
Run Growth: -1.2K
Growth Rate: -27.97%
Updated:July 17 2024
huggingface.co

Total runs: 3.9K
Run Growth: -788
Growth Rate: -20.07%
Updated:March 19 2026
huggingface.co

Total runs: 3.9K
Run Growth: -341
Growth Rate: -8.79%
Updated:June 23 2026
huggingface.co

Total runs: 3.3K
Run Growth: -998
Growth Rate: -30.33%
Updated:July 17 2024
huggingface.co

Total runs: 3.0K
Run Growth: 3.0K
Growth Rate: 100.00%
Updated:June 23 2026
huggingface.co

Total runs: 2.6K
Run Growth: -3.3K
Growth Rate: -127.96%
Updated:May 21 2026
huggingface.co

Total runs: 2.0K
Run Growth: -2.4K
Growth Rate: -119.11%
Updated:October 09 2025
huggingface.co

Total runs: 1.7K
Run Growth: -390
Growth Rate: -22.47%
Updated:April 11 2026