CanAiCode: leaderboard
Benchmarks smaller code-focused LLMs on text-to-code performance across multiple languages.
Metric: Junior-v2 Python Pass Rate (%). Source: huggingface.co. Status: saturated. 189 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Llama 3.1 8B Instruct | 100 |
| 2 | Claude 3.5 Sonnet (20240620) | 100 |
| 3 | GPT-4 Turbo | 100 |
| 4 | Claude 3 Opus (20240229) | 100 |
| 5 | Claude 3 Sonnet (20240229) | 100 |
| 6 | GPT-4 (0613) | 100 |
| 7 | GPT-4 Preview (1106) | 100 |
| 8 | Phi-3-small-8k-instruct | 100 |
| 9 | Phi-3-medium-128k-instruct | 100 |
| 10 | GPT-4 Preview (0125) | 100 |
| 11 | GPT-3.5 Turbo (0301) | 100 |
| 12 | Nxcode-CQ-7B-orpo | 100 |
| 13 | Gemma 2 27B (IT) | 98.9 |
| 14 | GPT-4o (2024-05-13) | 98.9 |
| 15 | Claude 3 Haiku (20240307) | 98.9 |
Interactive version: theaggregate.ai/benchmark?slug=canaicode · How It Works · Data refreshed daily, snapshot 2026-09-05.