whatmodel.sh / Dolly v2 12B
Dolly v2 12B
↗ share
+ contribute
Dolly v2 12B is an open-weights language model from Databricks, released April 2023. Its published size is 12B parameters. Licence: mit.
Catalogue identity notes
Comparison-only. Arena entry 'dolly-v2-12b' (text). LMArena's leaderboard table (arena-leaderboard space @6f2eeeb) labels 'dolly-v2-12b' as 'Dolly-V2-12B' (Databricks, MIT), linking https://huggingface.co/databricks/dolly-v2-12b. Parameters: nominal size from the model name. Released 2023-04: Databricks blog 'Free Dolly' (databricks.com/blog/2023/04/12/dolly-first-open-commercially-viable-instruction-tuned-llm), datePublished 2023-04-12. LMArena's table links huggingface.co/databricks/dolly-v2-12b (MIT), which no longer resolves; no current download verified.
$ how it compares
The LLMs ranked either side of Dolly v2 12B, out of 828 with a score, then more from Databricks.
| # | model | score | where | ||
|---|---|---|---|---|---|
| 803 |  BLOOM 3B | 18±1 | local | ||
| 804 |  FastChat-T5 3B | 17±2 | local | ||
| 805 | Dolly v2 12B | 17±2 | comparison only | ||
| 806 |  OPT-1.3B | 16±1 | local | ||
| 807 |  Gemma 3 270M IT | 15±4 | local | ||
| more from Databricks | |||||
| 509 | DBRX Instruct | 46±1 | local | ||
$ what it's good at
coding▁▁▃▃▅▇█▇▅▄▃▁ #516 of 524 * · 12% of its score · 1 test
Beats Molmo 7B-D (0924), DeepSeek R1 Distill Qwen 1.5B, Qwen3 0.6B here, though they rank higher overall.
chat & writing▃▁▁▃▅▄▆█▇▇▇▃ #461 of 463 · 81% of its score · 4 tests
Trails Gemma 3 270M IT here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| Polyrating rating (bias-aware re-fit of Chatbot Arena votes, mid-2024) | 856.0 ✎ | ▃▁▂▁▃▇█▆▇▄▂▂ beats 2/95 | 28% |
| LMArena (text) | 982.4 ✎ | ▃▁▁▂▃▃▃▆▆▇█▃ beats 2/379 | 28% |
| MT-Bench (GPT-4 single-answer grading, LMSYS runs) | 3.2800 MT-Bench score (1-10) ✎ | ▆▃▂▆▆▄██▂▄▄▆ beats 3/28 | 25% |
| MixInstruct GPT-Rank (LLM-Blender, mean rank among 11 instruction-tuned LLMs, ChatGPT pairwise judge) | 6.9000 mean rank (1-11) ✎ | █···▄··▄···▄ beats 2/4 | 0.0% |
$ where to run it
No verified API or download recorded yet · add one
$ how the score was worked out · Polyrating: A Cost-Effective… 28%, LMArena leaderboard dataset 28%, LMSYS Chatbot Arena leaderboa… 25% and 2 more
Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works
| source | results | share of the score |
|---|---|---|
 Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3) | 1 | 28% |
 LMArena leaderboard dataset | 1 | 28% |
 LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table | 1 | 25% |
 InstructEval (DeCLaRe Lab) README results table | 2 | 18% |
| LLM-Blender: Ensembling Large Language Models with Pairwise Ranking and Generative Fusion (arXiv:2306.02561v3) | 1 | 0% |
| prior · pull toward the middle | 1% | |
| weighted average | = 17.2 |
$ improve this page
Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.
data updated · sources and licences · json