Compare / nanbeige4-3b-thinking-2511 vs GPT-5.2
nanbeige4-3b-thinking-2511 vs GPT-5.2
11 results from 2 sources that measured both models. Each row is in the source's own units; there is no overall winner.
| Benchmark | Metric | nanbeige4-3b-thinking-2511 | GPT-5.2 |
|---|---|---|---|
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Irrelevance detection% correct | 83.1% | 87.3% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Memory% correct | 36.8% | 45.8% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Multi-turn tasks% correct | 51.1% | 43.8% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Overall accuracy% correct | 51.4% | 55.9% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Relevance detection% correct | 75.0% | 75.0% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Single-turn calls (curated)% correct | 81.6% | 81.8% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Single-turn calls (user-contributed)% correct | 79.4% | 70.4% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Web search% correct | 21.5% | 75.5% |
| UGI Leaderboardreported | Requested-length error% off the requested word count | 173.0% | 76.0% |
| UGI Leaderboardreported | Style adherencescore from 0 to 1 | 0.26 | 0.40 |
| UGI Leaderboardreported | Writing scorescore out of 100 | 9.8 | 49.4 |
Bold green marks the better value on that metric. Where a source reports ranges that overlap, the difference may not be meaningful; see the benchmark page for ranges.