1,330
Business, management and finance · rank 184 of 402
- Unit
- Arena rating, higher is better
- Range
- 1,300 to 1,360
- Sample
- 390 votes
- Configuration
- Gemma 3 4B
- Measured
- 25 Sep 2026
- Not shown
- Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.
1,274
Creative writing · rank 218 of 407
- Unit
- Arena rating, higher is better
- Range
- 1,252 to 1,297
- Sample
- 666 votes
- Configuration
- Gemma 3 4B
- Measured
- 25 Sep 2026
- Not shown
- Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.
1,260
Expert prompts · rank 238 of 359
- Unit
- Arena rating, higher is better
- Range
- 1,219 to 1,301
- Sample
- 208 votes
- Configuration
- Gemma 3 4B
- Measured
- 25 Sep 2026
- Not shown
- Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.
1,267
Instruction following · rank 269 of 409
- Unit
- Arena rating, higher is better
- Range
- 1,251 to 1,284
- Sample
- 1233 votes
- Configuration
- Gemma 3 4B
- Measured
- 25 Sep 2026
- Not shown
- Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.
1,303
Overall · rank 257 of 409
- Unit
- Arena rating, higher is better
- Range
- 1,294 to 1,313
- Sample
- 4171 votes
- Configuration
- Gemma 3 4B
- Measured
- 25 Sep 2026
- Not shown
- Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.
1,298
Writing, literature and language · rank 216 of 408
- Unit
- Arena rating, higher is better
- Range
- 1,281 to 1,315
- Sample
- 1080 votes
- Configuration
- Gemma 3 4B
- Measured
- 25 Sep 2026
- Not shown
- Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.
53.9%
Irrelevance detection · rank 96 of 109
- Unit
- % correct, higher is better
- Configuration
- gemma-3-4b-it
- Measured
- 16 Dec 2025
- Not shown
- Not general refusal or safety behaviour.
8.6%
Memory · rank 75 of 109
- Unit
- % correct, higher is better
- Configuration
- gemma-3-4b-it
- Measured
- 16 Dec 2025
- Not shown
- Not long-term personal memory in a product: sessions are BFCL's scripted ones.
0.4%
Multi-turn tasks · rank 103 of 109
- Unit
- % correct, higher is better
- Configuration
- gemma-3-4b-it
- Measured
- 16 Dec 2025
- Not shown
- Not open-ended agent work: the tools and tasks are BFCL's simulated APIs.
19.6%
Overall accuracy · rank 101 of 109
- Unit
- % correct, higher is better
- Configuration
- gemma-3-4b-it
- Measured
- 16 Dec 2025
- Not shown
- Not a neutral average: the weighting is BFCL's. Not reliability on your own tools: BFCL's functions and queries are a fixed test set, and a correct call is judged by its form, not by what it achieved.
100.0%
Relevance detection · rank 1 of 109
- Unit
- % correct, higher is better
- Configuration
- gemma-3-4b-it
- Measured
- 16 Dec 2025
- Not shown
- Not whether the call itself was right; only that one was attempted.
61.1%
Single-turn calls (curated) · rank 96 of 109
- Unit
- % correct, higher is better
- Configuration
- gemma-3-4b-it
- Measured
- 16 Dec 2025
- Not shown
- Not reliability on your own tools: BFCL's functions and queries are a fixed test set, and a correct call is judged by its form, not by what it achieved.
60.8%
Single-turn calls (user-contributed) · rank 83 of 109
- Unit
- % correct, higher is better
- Configuration
- gemma-3-4b-it
- Measured
- 16 Dec 2025
- Not shown
- Not reliability on your own tools: BFCL's functions and queries are a fixed test set, and a correct call is judged by its form, not by what it achieved.
1.0%
Web search · rank 81 of 109
- Unit
- % correct, higher is better
- Configuration
- gemma-3-4b-it
- Measured
- 16 Dec 2025
- Not shown
- Not general research quality: questions have short, checkable answers.
9.0%
Requested-length error · rank 67 of 370
- Unit
- % off the requested word count, lower is better
- Configuration
- Gemma 3 4B
- Measured
- 12 Sep 2025
- Not shown
- Not other format limits such as character counts or bullet counts.
0.29
Style adherence · rank 345 of 370
- Unit
- score from 0 to 1, higher is better
- Configuration
- Gemma 3 4B
- Measured
- 12 Sep 2025
- Not shown
- Not brand voice on your own examples: UGI's prompts are private and lean towards creative writing.
24.0
Writing score · rank 311 of 370
- Unit
- score out of 100, higher is better
- Configuration
- Gemma 3 4B
- Measured
- 12 Sep 2025
- Not shown
- Not business copy quality: UGI's prompts are private and lean towards creative writing, and models that often refuse get no score.
67.3%
Answer rate · rank 107 of 108
- Unit
- % of documents, higher is better
- Configuration
- Gemma 3 4B
- Measured
- 22 Sep 2026
- Not shown
- Not a quality score: a low rate usually means content filters were triggered, and hallucination rates are measured on answered documents only.
6.4%
Hallucination rate · rank 28 of 108
- Unit
- % of summaries, lower is better
- Configuration
- Gemma 3 4B
- Measured
- 22 Sep 2026
- Not shown
- Not errors in open questions or other tasks: only summarisation, judged by Vectara's own model (HHEM-2.3), not by people, on news-style documents rather than your data.
