BlogLLM Benchmark Scores as Published, Side by Side
MTOB: Results as Published
Every result the catalog holds for MTOB: 2 models from 1 developer, read from 1 page. Each number is as printed by the company named beside it; nothing is ranked or rescored. Models are listed by developer, newest first.
Area Long ContextMaintainer Garrett Tanzer and co-authorsIntroduced 2023Data version 2026103001, as of 2026-10-30
MTOB (Machine Translation from One Book) asks a model to translate between English and Kalamang, a language with fewer than 200 speakers and almost no text on the web. The material is one grammar book, a bilingual word list, and a small set of parallel sentences. After examples used in the book are removed, 500 sentences are split into 400 for training and 100 for test. The paper scores each direction with chrF.
MTOB (full book) eng->kgv (chrF)
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| Llama 4 MaverickMeta, released 2025-04-05 | 50.8 | Meta, Meta Llama 4 model cardchecked 2026-10-15 | - |
| Llama 4 ScoutMeta, released 2025-04-05 | 39.7 | Meta, Meta Llama 4 model cardchecked 2026-10-15 | - |
MTOB (full book) kgv->eng (chrF)
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| Llama 4 MaverickMeta, released 2025-04-05 | 46.7 | Meta, Meta Llama 4 model cardchecked 2026-10-15 | - |
| Llama 4 ScoutMeta, released 2025-04-05 | 36.3 | Meta, Meta Llama 4 model cardchecked 2026-10-15 | - |
MTOB (half book) eng->kgv (chrF)
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| Llama 4 MaverickMeta, released 2025-04-05 | 54 | Meta, Meta Llama 4 model cardchecked 2026-10-15 | - |
| Llama 4 ScoutMeta, released 2025-04-05 | 42.2 | Meta, Meta Llama 4 model cardchecked 2026-10-15 | - |
MTOB (half book) kgv->eng (chrF)
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| Llama 4 MaverickMeta, released 2025-04-05 | 46.4 | Meta, Meta Llama 4 model cardchecked 2026-10-15 | - |
| Llama 4 ScoutMeta, released 2025-04-05 | 36.6 | Meta, Meta Llama 4 model cardchecked 2026-10-15 | - |
Every number here is read from the same catalog file as the benchmark table, data version 2026103001, and is the number the table shows for that model and test. Where one page printed a model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.