OrionLLM/GRM2-3b
4B • Updated • 28 • 2
Orion LLM Labs
non-profit
AI & ML interests
Thinking, Agentic, and Research Purpose finetunning and training
Recent Activity
DedeProGames
updated a
model about 3 hours ago
Quantum Computing
#1 opened about 9 hours ago
by 
Reality123b
DedeProGames
updated a
collection 1 day ago
DedeProGames
posted an update 1 day ago
Post
87
Introducing GRM2, a powerful 3b parameter model designed for long-term reasoning and high performance in complex tasks.
Even with only 3b of parameters, it outperforms qwen3-32b in several benchmarks.
With only 3b of parameters, it can also generate large and complex code of over 1000 lines, use tools in a way comparable to large models, and is perfect for agentic tasks.
GRM2 is licensed under Apache 2.0, making it perfect as a FineTune base for other tasks.
OrionLLM/GRM2-3b
Even with only 3b of parameters, it outperforms qwen3-32b in several benchmarks.
With only 3b of parameters, it can also generate large and complex code of over 1000 lines, use tools in a way comparable to large models, and is perfect for agentic tasks.
GRM2 is licensed under Apache 2.0, making it perfect as a FineTune base for other tasks.
OrionLLM/GRM2-3b
DedeProGames
updated 2
models 1 day ago
DedeProGames
published a
model 2 days ago
DedeProGames
updated a
Space 2 days ago
DedeProGames
updated a
dataset 3 days ago
DedeProGames
published a
dataset 3 days ago
DedeProGames
updated a
model 3 days ago
DedeProGames
published a
model 3 days ago
DedeProGames
posted an update 3 days ago
Post
3736
Can small models program?
Although even if they are reasoning AIs, small AIs cannot create extensive and high-quality code, at least that's what is commonly thought.
We present OrionLLM/NanoCoder-0.6b, an AI with just 600 million parameters based on qwen3-0.6b and trained with the dataset nvidia/OpenCodeReasoning.
While not good at complex code, we observed a significant improvement in code generation (especially in Python code), demonstrating that, when trained correctly, small AIs can, in fact, program.
Although even if they are reasoning AIs, small AIs cannot create extensive and high-quality code, at least that's what is commonly thought.
We present OrionLLM/NanoCoder-0.6b, an AI with just 600 million parameters based on qwen3-0.6b and trained with the dataset nvidia/OpenCodeReasoning.
While not good at complex code, we observed a significant improvement in code generation (especially in Python code), demonstrating that, when trained correctly, small AIs can, in fact, program.
- 2 replies
DedeProGames
updated a
model 4 days ago
DedeProGames
posted an update 4 days ago
Post
2534
Introducing GRM Family, a family of fine-tuned small models from the Qwen2.5 family for Long Cot and General Reasoning and Agentic Tasks.
GRM is available in 7b and 1.5b parameter sizes, these models being significantly relevant for complex tasks or local inference agents.
OrionLLM/GRM-7b
OrionLLM/GRM-1.5b
GRM is available in 7b and 1.5b parameter sizes, these models being significantly relevant for complex tasks or local inference agents.
OrionLLM/GRM-7b
OrionLLM/GRM-1.5b
- 1 reply
DedeProGames
updated a
dataset 6 days ago
DedeProGames
published a
dataset 6 days ago
DedeProGames
updated a
dataset 6 days ago
DedeProGames
published a
dataset 6 days ago
DedeProGames
published a
model 7 days ago