Granite Code 34B Instruct: benchmark results
Provider: IBM. Released 2024-05-06. Access: Open.
Aggregate Rating 1390 ± 28, rank #6516 of 10916 rated models, from 107 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| EuroEval English NLU - SQuAD | 84.5 | Character F1 (%) | 83.1 |
| EuroEval German NLU - Germanquad | 59.52 | Character F1 (%) | 80 |
| EuroEval Spanish NLU - MLQA ES | 64.12 | Character F1 (%) | 77 |
| EuroEval Finnish NLU - Tydiqa FI | 66.84 | Character F1 (%) | 76.1 |
| EuroEval Portuguese NLU - HAREM | 46.63 | Micro F1 excluding MISC (%) | 69.4 |
| EuroEval Italian NLU - MultiNERD IT | 68.45 | Micro F1 excluding MISC (%) | 68.9 |
| EuroEval Portuguese NLU - MultiWikiQA PT | 71.92 | Character F1 (%) | 66.8 |
| EuroEval Norwegian NLU - NorNE NB | 66.01 | Micro F1 excluding MISC (%) | 58.5 |
| EuroEval Icelandic NLU - ScaLA IS | 7.58 | Matthews correlation coefficient (%) | 57.9 |
| EuroEval Norwegian NLU - Norquad | 65.16 | Character F1 (%) | 56.9 |
| EuroEval Spanish NLU - CoNLL ES | 62.39 | Micro F1 excluding MISC (%) | 56.6 |
| EuroEval Swedish NLU - SUC3 | 58 | Micro F1 excluding MISC (%) | 56 |
Interactive version: theaggregate.ai/model?slug=granite-code-34b-instruct · How It Works · Data refreshed daily, snapshot 2026-11-08.