Benchmark Leaderboards
Loading leaderboardsβ¦
Loading leaderboardsβ¦
Track which models hold the current SOTA on the benchmarks the field actually uses. Click a column header to sort; the leading row is highlighted.
FlowerTune Code track β federated fine-tuning of LLMs for code generation. Avg Score aggregates Pass@1-style accuracy over MBPP, HumanEval, and MultiPL-E (JavaScript, C++).
Metric: Avg Score β higher is better Β· Source β
| # β² | Model / Paper | Avg Score | Ξ vs SOTA | Links |
|---|---|---|---|---|
| 1 | ZeroOne.AI β Qwen3-8BSOTA | 65.27 | β | search β |
| 2 | Massimo R. Scamarcia β Qwen3-4B | 60.45 | -4.82 | search β |
| 3 | CAR@AIML β deepseek-coder-7b-instruct-v1.5 | 58.77 | -6.50 | search β |
| 4 | FL-finetune-JB-DC β Qwen2.5-Coder-7B-Instruct | 56.08 | -9.19 | search β |
| 5 | Massimo R. Scamarcia β Phi-4-mini-instruct | 49 | -16.27 | search β |
| 6 | CAR@AIML β starcoder2-7b | 44.08 | -21.19 | search β |
| 7 | Massimo R. Scamarcia β Qwen2.5-7B-Instruct | 34.4 | -30.87 | search β |
| 8 | CAR@AIML β CodeLlama-7b-hf | 33.78 | -31.49 | search β |