Benchmark Almanac

IPTC-test

Gemini 2.5 Flash Sep 2025 0.77 leads since September 2025
Not in model cards

Models choose one of 17 newsroom topics for each article. Human annotators labeled 1,129 articles in Catalan, Croatian, Greek, and Slovenian.

The authors report macro and micro F1 for 15 language model variants across all four languages. The authors share test articles on request.

Released
November 2024
Kind
Reading in many languages
Level
Everyday
Scored
Checked automatically
Answers
Short answer
Tasks
Public, plus a private test set
Measures
Capability
In and out
Text → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Strongest model tested
GPT-5 #71 on the capability index
Last model added
September 2025
Scores refreshed
November 20th, 2026
20
Entries on its own leaderboard, 11 models we track

Gemini 2.5 Flash leads with 0.77 among models released by September 2025 Gemini 2.5 Flash leads with 0.77 among models released by September 2025

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 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 Gemini 2.5 Flash Sep 2025 0.77
  2. 2 IPTC XLM-R classifier 0.74
  3. 3 GPT-4o Aug 2024 0.74
  4. 4 GPT-5 0.74
  5. 5 GPT-5 mini 0.73
  6. 6 GPT-4o mini 0.72
  7. 7 Mistral Medium 3.1 0.72
  8. 8 Gemma 3 27B 0.70
  9. 9 GPT-5 nano 0.69
  10. 10 Qwen3-32B 0.67
  11. 11 llama3.3:latest 0.63
  12. 12 GPT-3.5 Turbo Jan 2024 0.63
  13. 13 Llama 4 Scout 0.62
  14. 14 SVC 0.42
  15. 15 deepseek-r1:14b 0.39
  16. 16 GaMS-27B 0.24
  17. 17 GaMS-27B-quantized 0.23
  18. 18 COMPLEMENTNB 0.21
  19. 19 dummy-stratified 0.06
  20. 20 dummy-most_frequent 0.01

Source: Official IPTC-test leaderboard

#EntryScoreSettingDateSource
1 Gemini 2.5 Flash Sep 2025 0.77 Sep 2025 Official leaderboard
2 IPTC XLM-R classifier 0.74 Official leaderboard
3 GPT-4o Aug 2024 0.74 Aug 2024 Official leaderboard
4 GPT-5 0.74 Aug 2025 Official leaderboard
5 GPT-5 mini 0.73 Aug 2025 Official leaderboard
6 GPT-4o mini 0.72 Jul 2024 Official leaderboard
7 Mistral Medium 3.1 0.72 Aug 2025 Official leaderboard
8 Gemma 3 27B 0.70 Mar 2025 Official leaderboard
9 GPT-5 nano 0.69 Aug 2025 Official leaderboard
10 Qwen3-32B 0.67 Apr 2025 Official leaderboard
11 llama3.3:latest 0.63 Official leaderboard
12 GPT-3.5 Turbo Jan 2024 0.63 Jan 2024 Official leaderboard
13 Llama 4 Scout 0.62 Apr 2025 Official leaderboard
14 SVC 0.42 Official leaderboard
15 deepseek-r1:14b 0.39 Official leaderboard
16 GaMS-27B 0.24 Official leaderboard
17 GaMS-27B-quantized 0.23 Official leaderboard
18 COMPLEMENTNB 0.21 Official leaderboard
19 dummy-stratified 0.06 Official leaderboard
20 dummy-most_frequent 0.01 Official leaderboard
#ModelScoreSettingDateSource
1 Gemini 2.5 Flash Sep 2025 0.77 Sep 2025 Official leaderboard
2 GPT-4o Aug 2024 0.74 Aug 2024 Official leaderboard
3 GPT-5 0.74 Aug 2025 Official leaderboard
4 GPT-5 mini 0.73 Aug 2025 Official leaderboard
5 GPT-4o mini 0.72 Jul 2024 Official leaderboard
6 Mistral Medium 3.1 0.72 Aug 2025 Official leaderboard
7 Gemma 3 27B 0.70 Mar 2025 Official leaderboard
8 GPT-5 nano 0.69 Aug 2025 Official leaderboard
9 Qwen3-32B 0.67 Apr 2025 Official leaderboard
10 GPT-3.5 Turbo Jan 2024 0.63 Jan 2024 Official leaderboard
11 Llama 4 Scout 0.62 Apr 2025 Official leaderboard

Rankings by part

We read 8 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as hr macro F1, hr micro F1 and ca macro F1. Gemini 2.5 Flash leads 5 of them, and 2 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

hr macro F1

11 models we track have a score here.

  1. 1 GPT-5 ↑2 0.75
  2. 2 Gemini 2.5 Flash Sep 2025 ↓1 0.75
  3. 3 GPT-4o Aug 2024 ↓1 0.74
  4. 4 GPT-4o mini ↑1 0.73
  5. 5 GPT-5 mini ↓1 0.71
  6. 6 Mistral Medium 3.1 0.70
  7. 7 Gemma 3 27B 0.70
  8. 8 Qwen3-32B ↑1 0.68
  9. 9 GPT-5 nano ↓1 0.66
  10. 10 GPT-3.5 Turbo Jan 2024 0.59

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

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

OpenAI Google DeepMind
Own leaderboard Not tested since Sep 2025

Source: Official IPTC-test leaderboard

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
On the benchmark’s own leaderboard
0.65 0.70 0.75 0.80 2024 2025 0.65 0.70 0.75 0.80 2024 2025

Source: Official IPTC-test 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

8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 14 months ago, in September 2025, for Gemini 2.5 Flash. Every benchmark

Catalog record 2 pass
  • Pass Task access Public tasks, plus a private test set.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (20 entries).
  • Pass Models scored 11 models we track, plus 9 other entries on its leaderboard.
  • Warning Newest score The newest came 14 months ago, in September 2025, for Gemini 2.5 Flash.
  • Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5 (#71).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on github.com: 11 rows for models we track and 9 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