Benchmark Almanac

M-IFEval

o1-preview 86.6% leads since September 2024
Not in model cards

Models follow instructions in French, Japanese, and Spanish. Automatic checks test whether each answer satisfies the requested constraints.

The llm-jp-eval harness runs this evaluation with Japanese prompts.

Made by
Lightblue
Released
July 2024
Kind
Chat and instructions in many languages
Level
Everyday
Scored
Checked automatically
Answers
Open-ended
Tasks
All public
Measures
Capability
Language
French, Japanese, Spanish
In and out
Text → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Based on
IFEval
Strongest model tested
o1-mini #142 on the capability index
Last model added
September 2024
Scores refreshed
November 20th, 2026
17
Entries on its own leaderboard, 15 models we track

o1-preview leads with 86.6% among models released by September 2024 o1-preview leads with 86.6% among models released by September 2024

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.

The newest model with a score here came out in September 2024, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 o1-preview 86.6%
  2. 2 Claude 3 Opus 84.4%
  3. 3 Claude 3.5 Sonnet 84.2%
  4. 4 o1-mini 83.3%
  5. 5 GPT-4o Aug 2024 82.7%
  6. 6 GPT-4o mini 78.9%
  7. 7 Qwen2.5-32B GPTQ Int4 76.7%
  8. 8 Qwen2.5-14B GPTQ Int4 74.4%
  9. 9 Claude 3 Haiku 73%
  10. 10 Qwen2.5-7B GPTQ Int4 68.6%
  11. 11 Llama 3.1-8B AWQ Int4 63.8%
  12. 12 Qwen2.5-3B GPTQ Int4 59.8%
  13. 13 Mistral 7B v0.3 51.2%
  14. 14 Aya 23 8B 50.7%
  15. 15 Qwen2.5-1.5B GPTQ Int4 45.8%
  16. 16 DeepSeek LLM 7B Chat 39.7%
  17. 17 Qwen2.5-0.5B GPTQ Int4 30.1%

Source: Official M-IFEval leaderboard

#EntryScoreSettingDateSource
1 o1-preview 86.6% Sep 2024 Official leaderboard
2 Claude 3 Opus 84.4% Feb 2024 Official leaderboard
3 Claude 3.5 Sonnet 84.2% Jun 2024 Official leaderboard
4 o1-mini 83.3% Sep 2024 Official leaderboard
5 GPT-4o Aug 2024 82.7% Aug 2024 Official leaderboard
6 GPT-4o mini 78.9% Jul 2024 Official leaderboard
7 Qwen2.5-32B 76.7% GPTQ Int4 Sep 2024 Official leaderboard
8 Qwen2.5-14B 74.4% GPTQ Int4 Sep 2024 Official leaderboard
9 Claude 3 Haiku 73% Mar 2024 Official leaderboard
10 Qwen2.5-7B 68.6% GPTQ Int4 Sep 2024 Official leaderboard
11 Llama 3.1-8B 63.8% AWQ Int4 Jul 2024 Official leaderboard
12 Qwen2.5-3B 59.8% GPTQ Int4 Sep 2024 Official leaderboard
13 Mistral 7B v0.3 51.2% Sep 2023 Official leaderboard
14 Aya 23 8B 50.7% Official leaderboard
15 Qwen2.5-1.5B 45.8% GPTQ Int4 Sep 2024 Official leaderboard
16 DeepSeek LLM 7B Chat 39.7% Official leaderboard
17 Qwen2.5-0.5B 30.1% GPTQ Int4 Sep 2024 Official leaderboard
#ModelScoreSettingDateSource
1 o1-preview 86.6% Sep 2024 Official leaderboard
2 Claude 3 Opus 84.4% Feb 2024 Official leaderboard
3 Claude 3.5 Sonnet 84.2% Jun 2024 Official leaderboard
4 o1-mini 83.3% Sep 2024 Official leaderboard
5 GPT-4o Aug 2024 82.7% Aug 2024 Official leaderboard
6 GPT-4o mini 78.9% Jul 2024 Official leaderboard
7 Qwen2.5-32B 76.7% GPTQ Int4 Sep 2024 Official leaderboard
8 Qwen2.5-14B 74.4% GPTQ Int4 Sep 2024 Official leaderboard
9 Claude 3 Haiku 73% Mar 2024 Official leaderboard
10 Qwen2.5-7B 68.6% GPTQ Int4 Sep 2024 Official leaderboard
11 Llama 3.1-8B 63.8% AWQ Int4 Jul 2024 Official leaderboard
12 Qwen2.5-3B 59.8% GPTQ Int4 Sep 2024 Official leaderboard
13 Mistral 7B v0.3 51.2% Sep 2023 Official leaderboard
14 Qwen2.5-1.5B 45.8% GPTQ Int4 Sep 2024 Official leaderboard
15 Qwen2.5-0.5B 30.1% GPTQ Int4 Sep 2024 Official leaderboard

Rankings by part

We read 3 parts from the benchmark’s own leaderboard: Spanish, French and Japanese. Each ranks models on one part of the test. o1-preview leads 2 of them, and Claude 3.5 Sonnet leads the other one. Arrows show how far a model moved from its overall rank among the models we track.

Spanish

15 models we track have a score here.

  1. 1 o1-preview 92.7%
  2. 2 o1-mini ↑2 92%
  3. 3 Claude 3 Opus ↓1 90.5%
  4. 4 GPT-4o Aug 2024 ↑1 89.8%
  5. 5 Claude 3.5 Sonnet ↓2 87.6%
  6. 6 GPT-4o mini 85.4%
  7. 7 Qwen2.5-14B ↑1 83.2%
  8. 8 Qwen2.5-32B ↓1 82.5%
  9. 9 Llama 3.1-8B ↑2 80.3%
  10. 10 Claude 3 Haiku ↓1 78.8%

2 models have held number one on the official leaderboard since its release in July 2024

Lab claims count from the day the card came out, independent runs from the model’s release.

Anthropic OpenAI
Own leaderboard Passed 85% in Sep 2024, by o1-preview
o1-preview 86.6%

Source: Official M-IFEval leaderboard

O1-preview on the official leaderboard passed 85% in September 2024

That was 2 months after its release. By the same age, 10% of the benchmarks from 2024 had passed 85%. All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2024 Near the ceiling (85%) o1-preview Claude 3 Opus Mistral 7B v0.3 0% 20% 40% 60% 80% 100% 2024 Near the ceiling (85%) o1-preview Mistral 7B v0.3

Source: Official M-IFEval leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Multilingual

All 545 in Multilingual
  • MMMLU In use MMLU professionally translated into 14 languages. Multiple-choice accuracy. Claude Mythos Preview 92.7%
  • WMT24++ In use English-to-55-language translation of WMT24 test data. Scored with automatic metrics such as MetricX or xCOMET. K-EXAONE 90.5%
  • MMLU-ProX In use MMLU-Pro translated into up to 29 languages with expert review. Multiple-choice accuracy per language. Qwen3.7-Max 87%
  • CLIcK In use Models answer 1,995 multiple-choice questions on Korean language and culture. Questions come from official Korean exams and textbooks in eleven categories. Solar Open 2 90.7%
  • Global MMLU Lite 3.0 In use Models answer 400 human-translated MMLU questions per language in 23 languages. Version 3.0 adds Czech, Hungarian, Oriya, Slovak and Tajik and updates Italian. Accuracy. Claude Opus 5 95.7%
  • MMMB In use Models answer multiple-choice questions about images in six languages across 15 categories. The set has about 12,000 questions. LFM2.5-VL-3B 83%

Data audit

All 8 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
  • Pass Independent runs Scores from the official leaderboard (17 entries).
  • Pass Models scored 15 models we track, plus 2 other entries on its leaderboard.
  • Note Newest score The newest came 2 years ago, in September 2024, for Qwen2.5-0.5B. That’s expected for a saturated test.
  • Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is o1-mini (#142).
  • Note Headroom Saturated. The best is 86.6%, past 85% since September 2024.
Leaderboard import 3 pass
  • Pass Importer We read the board on github.com: 15 rows for models we track and 2 other entries at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check