allenai/tmax-9b
9B • Updated • 114 • 6
Tmax
Data and models associated with "Tmax: A simple recipe for terminal agents". paper: https://arxiv.org/abs/2606.23321
Note RLed model on top of Qwen 3.5 9B, with rollouts and model checkpoints through training.
allenai/tmax-27b
2.65M • Updated • 9 • 10Note RLed model on top of Qwen 3.6 27B.
allenai/tmax-4b
4B • Updated • 6 • 1Note RLed model on top of Qwen 3.5 4B.
allenai/tmax-2b
2B • Updated • 10 • 1Note RLed model on top of Qwen 3.5 2B.
allenai/tmax-sft-8b
8B • Updated • 7Note SFTed model on top of Qwen3 8B
allenai/tmax-8b
8B • Updated • 10Note RLed model on top of tmax-sft-8b
allenai/tmax-15k-open-instruct
Updated • 76 • 5Note RL data, formatted for use w/ open-instruct
allenai/TMax-15K
Viewer • Updated • 14.6k • 280 • 3Note Data for RL training for RL training for general use
allenai/tmax-sft
Updated • 43 • 1Note SFT data used for tmax-8b
allenai/TMax-SFT-16.5K
Viewer • Updated • 2.2k • 32Note SFT data, but with more details for general use below: data/models for various ablations.
-
allenai/tmax-sft-big
Viewer • Updated • 327k • 42 -
allenai/qwen35-9b-endless
9B • Updated • 18 -
allenai/open-instruct-endless-terminals
Updated • 34 -
allenai/qwen35-9b-terminaltraj
9B • Updated • 19 -
allenai/open-instruct-terminal-traj
Viewer • Updated • 5.65k • 36 -
allenai/qwen35-9b-termigen
9B • Updated • 22 -
allenai/open-instruct-termigen
Updated • 36 -
allenai/qwen35-9b-swesmith
9B • Updated • 18 -
allenai/open-instruct-swe-smith
Updated • 42 • 3 -
allenai/qwen35-9b-cli-gym
9B • Updated • 19 -
allenai/open-instruct-cli-gym
Updated • 35 -
allenai/qwen35-9b-openthoughts
9B • Updated • 14 • 2 -
allenai/open-instruct-openthoughts
Viewer • Updated • 728 • 36