GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
shawnmkim
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
shawnmkim
Model Tree
Qwen/Qwen2.5-Coder-7B
Input Modalities
Output Modalities
Supported Functionality
Held-out code_contests d6–9, 68 problems, execution pass@1:
| Method | pass@1 | Direction |
|---|---|---|
| Base (Qwen2.5-Coder-7B) | 5.9 % (4 / 68) | – |
| SFT (union of teachers) | 2.9 % | ↓ degrade |
| RLVR (GRPO, this adapter, peak checkpoints 250–750) | 8.8 % (6 / 68) | ↑ +49 % rel. |
| RLVR at step 1000 | 7.4 % (5 / 68) | ↑ |
Same data, opposite direction from SFT.
Learning curve:
| step | pass@1 (%) | solved / 68 |
|---|---|---|
| 0 | 5.9 | 4 |
| 100 | 5.9 | 4 |
| 250 | 8.8 | 6 |
| 500 | 8.8 | 6 |
| 750 | 8.8 | 6 |
| 1000 | 7.4 | 5 |
Differences within the 250–750 plateau are ±1 problem out of 68 (sampling noise); the robust signal is base → RLVR (4 → 5–6 / 68). We recommend checkpoint selection over final-step training.
This RLVR adapter is trained with NO teacher-output distillation. The reward signal is deterministic execution against hidden tests on publicly-licensed competition problems (code_contests, CC-BY-4.0). No raw outputs from Claude / GPT / Grok / Gemini were used as training targets.
This places the artifact outside the competition-scope restrictions of Anthropic / OpenAI / Google-Gemini / xAI (the paper only ranks those teachers as a scientific benchmark; it does not distill them to build a competitor).
python
import torchfrom transformers import AutoModelForCausalLM, AutoTokenizerfrom peft import PeftModel BASE = "Qwen/Qwen2.5-Coder-7B"LORA = "shawnmkim/qwen2.5-coder-7b-rlvr-compete-collab" tok = AutoTokenizer.from_pretrained(BASE)model = AutoModelForCausalLM.from_pretrained(BASE, torch_dtype=torch.bfloat16, device_map="auto")model = PeftModel.from_pretrained(model, LORA)model.eval() prompt = "Write a Python function that returns the n-th Fibonacci number."inputs = tok(prompt, return_tensors="pt").to(model.device)with torch.no_grad(): out = model.generate(**inputs, max_new_tokens=256, do_sample=False)print(tok.decode(out[0], skip_special_tokens=True))Qwen/Qwen2.5-Coder-7BGRPOTrainer), PEFT 0.19.1r=32, alpha=64, dropout 0q_proj, k_proj, v_proj, o_proj, up_proj, down_proj, gate_projnone, task type CAUSAL_LMtest_pass_fraction + small_format_bonustransformers 5.5 · torch 2.11 · cu130bash
git clone https://github.com/shawnkim678/compete-then-collaboratecd compete-then-collaboratepython reproduce.py --check-banks --selftestpython scripts/eval_code_students.py \ --base "Qwen/Qwen2.5-Coder-7B" \ --adapter "shawnmkim/qwen2.5-coder-7b-rlvr-compete-collab" \ --bank data/taskbank_contests_heldout.jsonl \ --label "grpo-7B [contests]"adapter_config.json — PEFT LoRA config (rank 32)adapter_model.safetensors — LoRA weights (~323 MB)chat_template.jinja — chat template inherited from basetokenizer.json / tokenizer_config.json — tokenizer copiesbibtex
@misc{kim2026compete, title = {Compete then Collaborate: Frontier AI Teachers Build a Verifiable Curriculum to Improve a Coding Student Beyond Imitation}, author = {Kim, Miseong (Shawn)}, year = {2026}, note = {Preprint, Genesis Cortex AI Inc.}, url = {https://github.com/shawnkim678/compete-then-collaborate}}Cite GRPO (method) and TRL (implementation):
bibtex
@article{shao2024deepseekmath, title = {DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models}, author = {Shao, Zhihong and Wang, Peiyi and Zhu, Qihao and others}, year = {2024}, eprint = {arXiv:2402.03300}} @misc{vonwerra2022trl, title = {{TRL: Transformer Reinforcement Learning}}, author = {von Werra, Leandro and Belkada, Younes and others}, year = {2020}, url = {https://github.com/huggingface/trl}}Following the accompanying paper, we thank Dr. Min Jang (POSTECH) — whose 1999 dissertation on ensemble learning with virtual data provided the conceptual lineage for the verifiable-curriculum framing — for reviewing an earlier draft and highlighting the engineering contributions of the reported system as its principal strengths (GB10 stack debugging, four-provider CLI orchestration, execution sandbox).
import torchfrom transformers import AutoModelForCausalLM, AutoTokenizerfrom peft import PeftModel BASE = "Qwen/Qwen2.5-Coder-7B"LORA = "shawnmkim/qwen2.5-coder-7b-rlvr-compete-collab" tok = AutoTokenizer.from_pretrained(BASE)model = AutoModelForCausalLM.from_pretrained(BASE, torch_dtype=torch.bfloat16, device_map="auto")model = PeftModel.from_pretrained(model, LORA)model.eval() prompt = "Write a Python function that returns the n-th Fibonacci number."inputs = tok(prompt, return_tensors="pt").to(model.device)with torch.no_grad(): out = model.generate(**inputs, max_new_tokens=256, do_sample=False)print(tok.decode(out[0], skip_special_tokens=True))git clone https://github.com/shawnkim678/compete-then-collaboratecd compete-then-collaboratepython reproduce.py --check-banks --selftestpython scripts/eval_code_students.py \ --base "Qwen/Qwen2.5-Coder-7B" \ --adapter "shawnmkim/qwen2.5-coder-7b-rlvr-compete-collab" \ --bank data/taskbank_contests_heldout.jsonl \ --label "grpo-7B [contests]"@misc{kim2026compete, title = {Compete then Collaborate: Frontier AI Teachers Build a Verifiable Curriculum to Improve a Coding Student Beyond Imitation}, author = {Kim, Miseong (Shawn)}, year = {2026}, note = {Preprint, Genesis Cortex AI Inc.}, url = {https://github.com/shawnkim678/compete-then-collaborate}}@article{shao2024deepseekmath, title = {DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models}, author = {Shao, Zhihong and Wang, Peiyi and Zhu, Qihao and others}, year = {2024}, eprint = {arXiv:2402.03300}} @misc{vonwerra2022trl, title = {{TRL: Transformer Reinforcement Learning}}, author = {von Werra, Leandro and Belkada, Younes and others}, year = {2020}, url = {https://github.com/huggingface/trl}}