Model card
phi-4.
microsoftopen-source14.7B params
§ 01 · Benchmarks
Every benchmark phi-4 has a recorded score for.
| # | Benchmark | Area · Task | Metric | Value | Rank | Date | Source |
|---|---|---|---|---|---|---|---|
| 01 | Open PL LLM Leaderboard | Natural Language Processing · Polish LLM General | belebele | 91.6% | #17 | — | source ↗ |
| 02 | Open PL LLM Leaderboard | Natural Language Processing · Polish LLM General | cbd | 37.3% | #29 | — | source ↗ |
| 03 | Open PL LLM Leaderboard | Natural Language Processing · Polish LLM General | polemo2-in | 86.0% | #34 | — | source ↗ |
| 04 | Polish EQ-Bench | Natural Language Processing · Polish Emotional Intelligence | eq-score | 59.1% | #34 | — | source ↗ |
| 05 | CPTU-Bench | Natural Language Processing · Polish Text Understanding | average | 3.3% | #47 | — | source ↗ |
| 06 | CPTU-Bench | Natural Language Processing · Polish Text Understanding | sentiment | 3.7% | #48 | — | source ↗ |
| 07 | CPTU-Bench | Natural Language Processing · Polish Text Understanding | tricky-questions | 2.7% | #48 | — | source ↗ |
| 08 | CPTU-Bench | Natural Language Processing · Polish Text Understanding | language-understanding | 3.5% | #48 | — | source ↗ |
| 09 | CPTU-Bench | Natural Language Processing · Polish Text Understanding | phraseology | 3.2% | #52 | — | source ↗ |
| 10 | Open PL LLM Leaderboard | Natural Language Processing · Polish LLM General | average | 62.6% | #55 | — | source ↗ |
| 11 | Open PL LLM Leaderboard | Natural Language Processing · Polish LLM General | dyk | 66.1% | #66 | — | source ↗ |
| 12 | Open PL LLM Leaderboard | Natural Language Processing · Polish LLM General | ppc | 76.3% | #91 | — | source ↗ |
| 13 | Open PL LLM Leaderboard | Natural Language Processing · Polish LLM General | polqa-open-book | 84.5% | #217 | — | source ↗ |
Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.
§ 02 · Strengths by area
Where phi-4 actually performs.
§ 04 · Related models
Other microsoft models scored on Codesota.
§ 05 · Sources & freshness
Where these numbers come from.
speakleash/open_pl_llm_leaderboard
7
results
SpeakLeash/CPTU-Bench
5
results
SpeakLeash/Polish-EQ-Bench
1
result
13 of 13 rows marked verified.