The questions a passage answers, in 324 languages, as many as you ask for.
MIST-QG-1.1 reads a passage and writes the questions it answers, in the passage’s own language or across languages, and you can say how many you want. It is a 4B model trained on 324 languages, built to make the (question, passage) pairs that embedders and rerankers learn from, in languages where those pairs barely exist.

Search and RAG systems learn from pairs of a question and the passage that answers it. In English there are millions of them. In most of the world’s languages there are almost none, so retrieval quietly works worse for most of the world.
MIST-QG-1.1 turns raw text in any of 324 languages into those pairs. It is small enough to run over millions of passages on one GPU, and its questions are written to find their own passage again, which is what a retrieval model needs to learn from.

46% of its questions are usable: on par with Qwen3.5-27B and ahead of Llama 3.3 70B. Only Gemma 4 31B, nearly eight times its size, scores higher.

English or French questions on passages in 153 other languages: 77% usable, the best of every model tested.

From 90 major languages down to 83 rare ones. In the rarest languages it scores 28%, against 15–25% for Qwen, gpt-oss and Llama up to 70B; Gemma 4 31B leads there at 32%.
Averaged over 238 test languages, 46% of its questions pass every check, on par with Qwen3.5-27B (47%) and ahead of Llama 3.3 70B, Qwen3.5-9B and gpt-oss-20b. Gemma 4 31B, nearly eight times its size, scores higher at 51%.
English or French questions on passages in 153 other languages: 77% usable, more than every model tested, Gemma 4 31B (67%) included. About 1.75 questions per passage, against one for MIST-QG-1.0, so more usable pairs from every passage.
Ask for “3 questions in Swahili” and get three. Cross-lingual, it returned exactly the number asked every time from 1 to 5; in the same language it treats the number as a maximum instead of padding.
Any of 324 languages into English or French, plus two directions 1.0 can’t do: English or French into 30 widely spoken languages (59% usable, 99.3% in the right language) and between any two of those 30.
from vllm import LLM, SamplingParams
def prompt(passage, lang, qlang, n=None):
ask = f"{n} {'question' if n == 1 else 'questions'}" if n else "Questions"
return (f"Passage ({lang}):\n{passage}\n\n"
f"{ask} in {qlang} that this passage answers, in your own words:\n")
llm = LLM(model="olaverse/MIST-QG-1.1", max_model_len=4096)
params = SamplingParams(max_tokens=320, temperature=0.3, top_p=0.9, stop=["\n\n"])
passage = "The Nile is the longest river in Africa. It flows north for about 6,650 km..."
prompts = [prompt(passage, "English", "English"), # the model decides how many
prompt(passage, "English", "Swahili", n=3)] # three questions in Swahili
for out in llm.generate(prompts, params):
print([l[2:].strip() for l in out.outputs[0].text.split("\n") if l.startswith("- ")])On the Olaverse API as mist-qg-1.1, with n: "auto" to let it decide how many. Would rather self-host a model that always lets the passage decide? Use MIST-QG-1.0.
| MIST-QG-1.1THIS PAGE | MIST-QG-1.0 → | |
|---|---|---|
| Parameters | 4B | 4B |
| Passage languages | 324 | 324 |
| How many questions | You choose 1 to 5: exact every time cross-lingually | The passage decides |
| Question languages | Own language, English, French, and 30 more | Own language, English or French |
| English or French → 30 languages | 59% usable, 99.3% in the right language | Not supported |
| Between two of those 30 languages | 47% usable | Not supported |
| Cross-lingual questions per passage | About 1.75 | About 1 |
| Questions that find their own passage (R@1) | 0.701, the best tested | 0.696 |
| Best for | A set number of questions, in more languages and directions | Letting the passage decide how many |
| On the Olaverse API | Yes | No, self-host only |
Generate the questions a passage answers, in 324 languages, in its own language or across languages, for quizzes, reading comprehension, help-centre FAQs and assessment banks.
/ ML engineeringGenerate verified (question, passage) pairs from your own corpus in 324 languages, or across languages, and fine-tune retrievers and rerankers on data that matches your domain.
typed-decisions 0.723 · 4.5B
PurpleMISTtyped-decisions 0.720 · African TD 0.668
PurpleMISTtyped-decisions 0.601 · 1.9B
MIST-Extract91.8% docs fully right · 4B
MIST-Extract9.7 docs/s · 2B
LID-6080.867 acc · 140M
LID-60837 MB · 6,800 texts/s
DiacNetDER 0.0117 · 582M
DiacNetDER 0.0141 · 300M
DiacNetDER 0.0130 · 37.9M