New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
Multi-IF benchmark
AI model leaderboard for the Multi-IF benchmark. Compare how large language models score on Multi-IF, see the full ranking, and understand what this AI benchmark measures. Qwen3-14B currently leads with 74.8. Multilingual instruction-following benchmark — IFEval-style prompts in 8 languages.
Leaderboard
| # | Model | Organization | Score | Variant | Source |
|---|---|---|---|---|---|
| #1 | Qwen3-14B | Qwen | 74.8 | thinking | official ↗ |
| #2 | Qwen3-32B | Qwen | 73 | thinking | official ↗ |
| #3 | Qwen3-14B | Qwen | 72.9 | non-thinking | official ↗ |
| #4 | Qwen3-30B-A3B | Qwen | 72.2 | thinking | official ↗ |
| #5 | Qwen3-235B-A22B | Qwen | 71.9 | thinking | official ↗ |
| #6 | Qwen3-8B | Qwen | 71.2 | thinking | official ↗ |
| #7 | Qwen3-30B-A3B | Qwen | 70.8 | non-thinking | official ↗ |
| #8 | Qwen3-32B | Qwen | 70.7 | non-thinking | official ↗ |
| #9 | Qwen3-235B-A22B | Qwen | 70.2 | non-thinking | official ↗ |
| #10 | Qwen3-8B | Qwen | 69.2 | non-thinking | official ↗ |
| #11 | Qwen3-4B | Qwen | 66.3 | thinking | official ↗ |
| #12 | Qwen3-4B | Qwen | 61.3 | non-thinking | official ↗ |
| #13 | Qwen3-1.7B | Qwen | 51.2 | thinking | official ↗ |
| #14 | Qwen3-1.7B | Qwen | 44.7 | non-thinking | official ↗ |
| #15 | Qwen3-0.6B | Qwen | 36.1 | thinking | official ↗ |
| #16 | Qwen3-0.6B | Qwen | 33.3 | non-thinking | official ↗ |
Frequently asked questions about Multi-IF
What is the Multi-IF benchmark?
Multilingual instruction-following benchmark — IFEval-style prompts in 8 languages.
How is the Multi-IF benchmark scored?
Multi-IF is scored using the accuracy metric, where a higher score is better. GenAIList aggregates reported scores from model providers and papers into a single ranked leaderboard.
Which AI model scores highest on Multi-IF?
As of the latest reported scores on GenAIList, Qwen3-14B achieves the highest result on Multi-IF with a score of 74.8.
Is a higher Multi-IF score better?
Yes. On Multi-IF a higher score indicates better performance, so models near the top of the leaderboard are the strongest.