Benchmark Almanac

INJONGO

AfroXLMR-76L 94.4% tops its leaderboard GPT-4o Aug 2024 65.1% leads the models we track
Not in model cards

Models identify a speaker’s intent and extract information from utterances in 16 African languages.

AfroBench uses the intent detection component.

Made by
Hao Yu and colleagues
Released
February 2025
Kind
Reading in many languages
Level
Everyday
Scored
Checked automatically
Answers
Short answer
Tasks
All public
Measures
Capability
Language
Many languages
In and out
Text → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
In views
Many languages
Strongest model tested
GPT-4o Aug 2024 #167 on the capability index
Last model added
December 2024
Scores refreshed
November 20th, 2026
15
Entries in its paper, 5 models we track

AfroXLMR-76L tops the leaderboard with 94.4% GPT-4o leads with 65.1% among models released by December 2024

Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. The headline is mean intent detection accuracy across the 16 African languages. 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. The best model we track is GPT-4o, at 65.1%. This view shows only models we track, each at its best.

The newest model with a score here came out in December 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 AfroXLMR-76L Multilingual training 94.4%
  2. 2 AfroXLMR 76L In-language training 93.7%
  3. 3 NLLB LLM2Vec In-language training 93.4%
  4. 4 AfroXLMR In-language training 92.2%
  5. 5 AfriTeVa V2 (T5) In-language training 91.7%
  6. 6 AfriBERTa V2 In-language training 88.6%
  7. 7 mT5-Large In-language training 87.7%
  8. 8 XLM-RoBERTa In-language training 84.5%
  9. 9 GPT-4o Aug 2024, Zero-shot 65.1%
  10. 10 Gemini 1.5 Pro (snapshot unspecified) Zero-shot 59.1%
  11. 11 Aya-101 13B Zero-shot 42.4%
  12. 12 Llama 3.3 70B Instruct Zero-shot 38.5%
  13. 13 Gemma 2 27B Zero-shot 33%
  14. 14 Gemma 2 9B Zero-shot 29.2%
  15. 15 Llama 3.1-8B Zero-shot 3.90%

Source: The INJONGO paper

#EntryScoreSettingDateSource
1 AfroXLMR-76L 94.4% Multilingual training Official leaderboard
2 AfroXLMR 76L 93.7% In-language training Official leaderboard
3 NLLB LLM2Vec 93.4% In-language training Official leaderboard
4 AfroXLMR 92.2% In-language training Official leaderboard
5 AfriTeVa V2 (T5) 91.7% In-language training Official leaderboard
6 AfriBERTa V2 88.6% In-language training Official leaderboard
7 mT5-Large 87.7% In-language training Official leaderboard
8 XLM-RoBERTa 84.5% In-language training Official leaderboard
9 GPT-4o Aug 2024 65.1% Zero-shot Aug 2024 Official leaderboard
10 Gemini 1.5 Pro (snapshot unspecified) 59.1% Zero-shot Official leaderboard
11 Aya-101 13B 42.4% Zero-shot Official leaderboard
12 Llama 3.3 70B Instruct 38.5% Zero-shot Dec 2024 Official leaderboard
13 Gemma 2 27B 33% Zero-shot Jun 2024 Official leaderboard
14 Gemma 2 9B 29.2% Zero-shot Jun 2024 Official leaderboard
15 Llama 3.1-8B 3.90% Zero-shot Jul 2024 Official leaderboard
#ModelScoreSettingDateSource
1 GPT-4o Aug 2024 65.1% Zero-shot Aug 2024 Official leaderboard
2 Llama 3.3 70B Instruct 38.5% Zero-shot Dec 2024 Official leaderboard
3 Gemma 2 27B 33% Zero-shot Jun 2024 Official leaderboard
4 Gemma 2 9B 29.2% Zero-shot Jun 2024 Official leaderboard
5 Llama 3.1-8B 3.90% Zero-shot Jul 2024 Official leaderboard

Rankings by part

We read 35 parts from the benchmark’s paper in 2 groups: intent detection and slot filling. Each ranks models on one part of the test. GPT-4o leads 32 of them, and Llama 3.3 70B Instruct leads the other 3. Arrows show how far a model moved from its overall rank among the models we track.

eng

5 models we track have a score here.

  1. 1 Llama 3.3 70B Instruct ↑1 81.1%
  2. 2 GPT-4o Aug 2024 ↓1 80.9%
  3. 3 Gemma 2 27B 79.5%
  4. 4 Gemma 2 9B 77.6%
  5. 5 Llama 3.1-8B 27.6%

Number one over time

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

OpenAI
Benchmark paper Not tested since Feb 2025
GPT-4o 65.1%

Source: The INJONGO paper

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
In the benchmark’s paper Best so far
0% 20% 40% 60% 80% 100% 2024 2025 Near the ceiling (85%) GPT-4o Gemma 2 27B 0% 20% 40% 60% 80% 100% 2024 2025 Near the ceiling (85%) GPT-4o Gemma 2 27B

Source: Official INJONGO 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

9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 23 months ago, in December 2024, for Llama 3.3 70B Instruct. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
  • Pass Independent runs Scores from the benchmark’s paper (15 entries).
  • Pass Models scored 5 models we track, plus 10 other entries on its leaderboard.
  • Warning Newest score The newest came 23 months ago, in December 2024, for Llama 3.3 70B Instruct.
  • Note Frontier models None of the top 10 has a score. The strongest tested is GPT-4o (#167).
  • Pass Headroom The best is 65.1%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on aclanthology.org: 5 rows for models we track and 10 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