BlogLLM Benchmark Scores as Published, Side by Side
AdvancedIF: Results as Published
Every result the catalog holds for AdvancedIF: 2 models from 2 developers, 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 Instruction FollowingMaintainer Yun He, Wenzhe Li, Hejia Zhang and co-authors (code published on GitHub under facebookresearch)Introduced 2025Data version 2026103001, as of 2026-10-30
AdvancedIF is a benchmark of over 1,600 prompts with expert-curated rubrics, testing whether a model follows complex, multi-turn and system-level instructions. The model is judged against each rubric item, and the score reported is the rubric-level score.
AdvancedIF (rubric-level score)
rubric-level score
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| Claude Sonnet 4.6Anthropic, released 2026-02-17 | 86 | Microsoft, Microsoft AI: MAI-Thinking-1 technical reportchecked 2026-10-19 | Effort: max |
| MAI-Thinking-1Microsoft, released 2026-08-12 | 85 | Microsoft, Microsoft AI: MAI-Thinking-1 technical reportchecked 2026-10-19 | - |
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.