IndicIFEval
Models follow instructions in 14 Indic languages. Rule-based checks test constraints such as required words, paragraph structure and word counts.
The February 2026 paper compares Gemini, GPT, Aya, Qwen and other models. Gemini 3 Pro led on translated prompts with 90.0% average loose accuracy.
- Made by
- AI4Bharat
- Released
- February 2026
- Kind
- Chat and instructions in many languages
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Open-ended
- Tasks
- All public
- Measures
- Capability
- Language
- Many languages
- Interaction
- Single turn
- Items from
- Translated
- Refresh
- One fixed set
- Run by
- Anyone
- Based on
- IFEval
- In views
- Many languages
- Strongest model tested
-
Gemini 3 Pro #56 on the capability index
- Last model added
- December 2025
- Scores refreshed
- November 20th, 2026
GPT-5 leads with 82.4% among models released by December 2025 GPT-5 leads with 82.4% among models released by December 2025
Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. The paper gives no combined score. The headline is the plain mean of translated and grounded prompt accuracy. 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 December 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 82.4 GPT-5
- 77.8 GPT-5 mini
- 76.3
Gemini 3 Pro
- 74.8
Gemini 3 Flash
- 70.5
Qwen3-14B
- 69.7
Gemma 3 27B
- 69.4
Qwen3-32B
- 67.8
Gemma 3 12B
- 64.7
Qwen3-8B
- 60.1
Gemma 3 4B
- 58.1
Qwen3-4B
- 51.1 meta-llama/Llama-3.1-8B-Instruct
- 45.2
Qwen3-1.7B
- 44.3
Aya Expanse 32B
- 44.1 meta-llama/Llama-3.2-3B-Instruct
- 37.6
Gemma 3 1B
- 37.0
Qwen3-0.6B
- 35.5 meta-llama/Llama-3.2-1B-Instruct
- 32.9
Aya Expanse 8B
- 1 GPT-5 82.4%
- 2 GPT-5 mini 77.8%
- 3
Gemini 3 Pro 76.3%
- 4
Gemini 3 Flash 74.8%
- 5
Qwen3-14B 70.5%
- 6
Gemma 3 27B 69.7%
- 7
Qwen3-32B 69.4%
- 8
Gemma 3 12B 67.8%
- 9
Qwen3-8B 64.7%
- 10
Gemma 3 4B 60.1%
- 11
Qwen3-4B 58.1%
- 12 meta-llama/Llama-3.1-8B-Instruct 51.1%
- 13
Qwen3-1.7B 45.2%
- 14
Aya Expanse 32B 44.3%
- 15 meta-llama/Llama-3.2-3B-Instruct 44.1%
- 16
Gemma 3 1B 37.6%
- 17
Qwen3-0.6B 37%
- 18 meta-llama/Llama-3.2-1B-Instruct 35.5%
- 19
Aya Expanse 8B 32.9%
Source: The IndicIFEval paper
- 1 GPT-5 82.4%
- 2 GPT-5 mini 77.8%
- 3
Gemini 3 Pro 76.3%
- 4
Gemini 3 Flash 74.8%
- 5
Qwen3-14B 70.5%
- 6
Gemma 3 27B 69.7%
- 7
Qwen3-32B 69.4%
- 8
Gemma 3 12B 67.8%
- 9
Qwen3-8B 64.7%
- 10
Gemma 3 4B 60.1%
- 11
Qwen3-4B 58.1%
- 12
Qwen3-1.7B 45.2%
- 13
Aya Expanse 32B 44.3%
- 14
Gemma 3 1B 37.6%
- 15
Qwen3-0.6B 37%
- 16
Aya Expanse 8B 32.9%
Source: The IndicIFEval paper
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-5 | 82.4% | Aug 2025 | Official leaderboard | |
| 2 | GPT-5 mini | 77.8% | Aug 2025 | Official leaderboard | |
| 3 | | 76.3% | Nov 2025 | Official leaderboard | |
| 4 | | 74.8% | Dec 2025 | Official leaderboard | |
| 5 | | 70.5% | Apr 2025 | Official leaderboard | |
| 6 | | 69.7% | Mar 2025 | Official leaderboard | |
| 7 | | 69.4% | Apr 2025 | Official leaderboard | |
| 8 | | 67.8% | Mar 2025 | Official leaderboard | |
| 9 | | 64.7% | Apr 2025 | Official leaderboard | |
| 10 | | 60.1% | Mar 2025 | Official leaderboard | |
| 11 | | 58.1% | Apr 2025 | Official leaderboard | |
| 12 | meta-llama/Llama-3.1-8B-Instruct | 51.1% | Official leaderboard | ||
| 13 | | 45.2% | Apr 2025 | Official leaderboard | |
| 14 | | 44.3% | Oct 2024 | Official leaderboard | |
| 15 | meta-llama/Llama-3.2-3B-Instruct | 44.1% | Official leaderboard | ||
| 16 | | 37.6% | Mar 2025 | Official leaderboard | |
| 17 | | 37% | Apr 2025 | Official leaderboard | |
| 18 | meta-llama/Llama-3.2-1B-Instruct | 35.5% | Official leaderboard | ||
| 19 | | 32.9% | Oct 2024 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-5 | 82.4% | Aug 2025 | Official leaderboard | |
| 2 | GPT-5 mini | 77.8% | Aug 2025 | Official leaderboard | |
| 3 | | 76.3% | Nov 2025 | Official leaderboard | |
| 4 | | 74.8% | Dec 2025 | Official leaderboard | |
| 5 | | 70.5% | Apr 2025 | Official leaderboard | |
| 6 | | 69.7% | Mar 2025 | Official leaderboard | |
| 7 | | 69.4% | Apr 2025 | Official leaderboard | |
| 8 | | 67.8% | Mar 2025 | Official leaderboard | |
| 9 | | 64.7% | Apr 2025 | Official leaderboard | |
| 10 | | 60.1% | Mar 2025 | Official leaderboard | |
| 11 | | 58.1% | Apr 2025 | Official leaderboard | |
| 12 | | 45.2% | Apr 2025 | Official leaderboard | |
| 13 | | 44.3% | Oct 2024 | Official leaderboard | |
| 14 | | 37.6% | Mar 2025 | Official leaderboard | |
| 15 | | 37% | Apr 2025 | Official leaderboard | |
| 16 | | 32.9% | Oct 2024 | Official leaderboard |
Rankings by part
We read 31 parts from the benchmark’s paper, each ranking models on one part of the test, such as IndicIFEval-Trans, IndicIFEval-Trans as and IndicIFEval-Trans bn. Gemini 3 Pro leads 11 of them, and 4 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
IndicIFEval-Trans
16 models we track have a score here.
- 1
Gemini 3 Pro ↑2 90%
- 2 GPT-5 ↓1 88.9%
- 3
Gemini 3 Flash ↑1 87.5%
- 4 GPT-5 mini ↓2 79.8%
- 5
Gemma 3 27B ↑1 76.1%
- 6
Gemma 3 12B ↑2 73.6%
- 7
Qwen3-32B 72.8%
- 8
Qwen3-14B ↓3 71.1%
- 9
Qwen3-8B 62.7%
- 10
Gemma 3 4B 60.1%
IndicIFEval-Trans as
16 models we track have a score here.
- 1
Gemini 3 Flash ↑3 90%
- 2
Gemini 3 Pro ↑1 89.4%
- 3 GPT-5 ↓2 88.8%
- 4 GPT-5 mini ↓2 77.9%
- 5
Gemma 3 27B ↑1 72.3%
- 6
Gemma 3 12B ↑2 71.7%
- 7
Qwen3-32B 71.7%
- 8
Qwen3-14B ↓3 69.2%
- 9
Qwen3-8B 59.5%
- 10
Qwen3-4B ↑1 54.2%
IndicIFEval-Trans bn
16 models we track have a score here.
- 1
Gemini 3 Pro ↑2 92.5%
- 2 GPT-5 ↓1 90.7%
- 3
Gemini 3 Flash ↑1 90%
- 4 GPT-5 mini ↓2 86%
- 5
Gemma 3 27B ↑1 81%
- 6
Gemma 3 12B ↑2 78.8%
- 7
Gemma 3 4B ↑3 73.8%
- 8
Qwen3-32B ↓1 73.8%
- 9
Qwen3-14B ↓4 72.6%
- 10
Qwen3-8B ↓1 67.3%
IndicIFEval-Trans gu
16 models we track have a score here.
- 1
Gemini 3 Pro ↑2 91.3%
- 2
Gemini 3 Flash ↑2 89.7%
- 3 GPT-5 ↓2 89.4%
- 4 GPT-5 mini ↓2 84.4%
- 5
Gemma 3 27B ↑1 80.4%
- 6
Gemma 3 12B ↑2 77%
- 7
Qwen3-32B 74.8%
- 8
Qwen3-14B ↓3 72.9%
- 9
Gemma 3 4B ↑1 65.4%
- 10
Qwen3-8B ↓1 63.9%
IndicIFEval-Trans hi
16 models we track have a score here.
- 1 GPT-5 92.8%
- 2
Gemini 3 Pro ↑1 91%
- 3
Gemini 3 Flash ↑1 89.7%
- 4 GPT-5 mini ↓2 88.8%
- 5
Gemma 3 27B ↑1 85%
- 6
Gemma 3 12B ↑2 81.9%
- 7
Qwen3-14B ↓2 78.8%
- 8
Qwen3-32B ↓1 78.5%
- 9
Gemma 3 4B ↑1 75.1%
- 10
Qwen3-8B ↓1 74.1%
IndicIFEval-Trans kn
16 models we track have a score here.
- 1
Gemini 3 Pro ↑2 91.3%
- 2 GPT-5 ↓1 89.1%
- 3
Gemini 3 Flash ↑1 86.3%
- 4
Gemma 3 27B ↑2 76.3%
- 5
Qwen3-14B 75.1%
- 6
Qwen3-32B ↑1 74.1%
- 7
Gemma 3 12B ↑1 73.8%
- 8 GPT-5 mini ↓6 72.9%
- 9
Qwen3-8B 67.9%
- 10
Qwen3-4B ↑1 54.2%
IndicIFEval-Trans ml
16 models we track have a score here.
- 1 GPT-5 90%
- 2
Gemini 3 Pro ↑1 86.9%
- 3
Gemini 3 Flash ↑1 84.7%
- 4
Gemma 3 27B ↑2 77.3%
- 5 GPT-5 mini ↓3 76.3%
- 6
Qwen3-32B ↑1 73.5%
- 7
Gemma 3 12B ↑1 72.6%
- 8
Qwen3-14B ↓3 69.5%
- 9
Qwen3-8B 61.7%
- 10
Gemma 3 4B 56.7%
IndicIFEval-Trans mr
16 models we track have a score here.
- 1
Gemini 3 Pro ↑2 91.3%
- 2 GPT-5 ↓1 89.7%
- 3
Gemini 3 Flash ↑1 88.8%
- 4 GPT-5 mini ↓2 81.6%
- 5
Gemma 3 27B ↑1 77.6%
- 6
Gemma 3 12B ↑2 77.3%
- 7
Qwen3-32B 75.7%
- 8
Qwen3-14B ↓3 70.4%
- 9
Gemma 3 4B ↑1 68.2%
- 10
Qwen3-8B ↓1 66.7%
IndicIFEval-Trans ne
16 models we track have a score here.
- 1
Gemini 3 Pro ↑2 92.2%
- 2 GPT-5 ↓1 91%
- 3
Gemini 3 Flash ↑1 89.7%
- 4 GPT-5 mini ↓2 82.9%
- 5
Gemma 3 27B ↑1 79.4%
- 6
Qwen3-32B ↑1 73.5%
- 7
Gemma 3 12B ↑1 72.9%
- 8
Qwen3-14B ↓3 72.3%
- 9
Gemma 3 4B ↑1 65.4%
- 10
Qwen3-8B ↓1 64.8%
IndicIFEval-Trans or
16 models we track have a score here.
- 1 GPT-5 89.1%
- 2
Gemini 3 Flash ↑2 86.9%
- 3
Gemini 3 Pro 86%
- 4 GPT-5 mini ↓2 75.4%
- 5
Qwen3-14B 74.5%
- 6
Qwen3-32B ↑1 70.4%
- 7
Gemma 3 27B ↓1 69.2%
- 8
Gemma 3 12B 65.7%
- 9
Qwen3-8B 52.6%
- 10
Qwen3-4B ↑1 47.4%
IndicIFEval-Trans pa
16 models we track have a score here.
- 1
Gemini 3 Pro ↑2 93.2%
- 2 GPT-5 ↓1 90.3%
- 3
Gemini 3 Flash ↑1 89.4%
- 4 GPT-5 mini ↓2 84.1%
- 5
Gemma 3 27B ↑1 75.7%
- 6
Gemma 3 12B ↑2 71.3%
- 7
Qwen3-14B ↓2 70.1%
- 8
Qwen3-32B ↓1 67.6%
- 9
Qwen3-8B 57.6%
- 10
Gemma 3 4B 50.2%
IndicIFEval-Trans sa
16 models we track have a score here.
- 1
Gemini 3 Pro ↑2 86.3%
- 2 GPT-5 ↓1 80.4%
- 3
Gemini 3 Flash ↑1 79.4%
- 4 GPT-5 mini ↓2 70.1%
- 5
Gemma 3 27B ↑1 64.8%
- 6
Qwen3-32B ↑1 64.8%
- 7
Qwen3-14B ↓2 64.5%
- 8
Gemma 3 12B 63.5%
- 9
Qwen3-8B 54.5%
- 10
Qwen3-4B ↑1 48.3%
IndicIFEval-Trans ta
16 models we track have a score here.
- 1
Gemini 3 Pro ↑2 89.7%
- 2 GPT-5 ↓1 86.9%
- 3
Gemini 3 Flash ↑1 85.7%
- 4 GPT-5 mini ↓2 75.1%
- 5
Gemma 3 27B ↑1 73.2%
- 6
Qwen3-32B ↑1 73.2%
- 7
Gemma 3 12B ↑1 71.7%
- 8
Gemma 3 4B ↑2 67.6%
- 9
Qwen3-14B ↓4 65.4%
- 10
Qwen3-8B ↓1 60.8%
IndicIFEval-Trans te
16 models we track have a score here.
- 1
Gemini 3 Pro ↑2 90.3%
- 2
Gemini 3 Flash ↑2 89.4%
- 3 GPT-5 ↓2 88.5%
- 4 GPT-5 mini ↓2 80.1%
- 5
Gemma 3 27B ↑1 76%
- 6
Gemma 3 12B ↑2 74.5%
- 7
Qwen3-32B 72%
- 8
Qwen3-14B ↓3 69.8%
- 9
Gemma 3 4B ↑1 64.8%
- 10
Qwen3-8B ↓1 62%
IndicIFEval-Trans ur
16 models we track have a score here.
- 1
Gemini 3 Pro ↑2 89.1%
- 2 GPT-5 ↓1 87.5%
- 3
Gemini 3 Flash ↑1 85.4%
- 4 GPT-5 mini ↓2 81.3%
- 5
Gemma 3 12B ↑3 77.9%
- 6
Gemma 3 27B 77.6%
- 7
Qwen3-32B 75.4%
- 8
Qwen3-14B ↓3 70.4%
- 9
Gemma 3 4B ↑1 67.3%
- 10
Qwen3-8B ↓1 64.8%
IndicIFEval-Trans en
16 models we track have a score here.
- 1
Gemini 3 Flash ↑3 96.3%
- 2
Gemini 3 Pro ↑1 95.3%
- 3 GPT-5 ↓2 95%
- 4 GPT-5 mini ↓2 91.9%
- 5
Gemma 3 27B ↑1 88.2%
- 6
Qwen3-8B ↑3 88.2%
- 7
Qwen3-14B ↓2 87.8%
- 8
Qwen3-32B ↓1 87.2%
- 9
Gemma 3 12B ↓1 86%
- 10
Qwen3-4B ↑1 85%
IndicIFEval-Ground
16 models we track have a score here.
- 1 GPT-5 75.9%
- 2 GPT-5 mini 75.8%
- 3
Qwen3-14B ↑2 70%
- 4
Qwen3-8B ↑5 66.6%
- 5
Qwen3-32B ↑2 66%
- 6
Gemma 3 27B 63.3%
- 7
Gemini 3 Pro ↓4 62.6%
- 8
Gemini 3 Flash ↓4 62.1%
- 9
Gemma 3 12B ↓1 62%
- 10
Qwen3-4B ↑1 62%
IndicIFEval-Ground as
16 models we track have a score here.
- 1 GPT-5 mini ↑1 75.5%
- 2
Qwen3-14B ↑3 74.3%
- 3
Qwen3-32B ↑4 73.2%
- 4
Qwen3-8B ↑5 73.2%
- 5 GPT-5 ↓4 73.2%
- 6
Qwen3-4B ↑5 71.2%
- 7
Gemma 3 27B ↓1 69.3%
- 8
Gemini 3 Flash ↓4 64.6%
- 9
Gemini 3 Pro ↓6 62.3%
- 10
Gemma 3 4B 61.5%
IndicIFEval-Ground bn
16 models we track have a score here.
- 1 GPT-5 72.6%
- 2
Qwen3-14B ↑3 72.1%
- 3 GPT-5 mini ↓1 71.2%
- 4
Qwen3-32B ↑3 70.2%
- 5
Qwen3-8B ↑4 70.2%
- 6
Qwen3-4B ↑5 69.4%
- 7
Gemini 3 Flash ↓3 67%
- 8
Gemini 3 Pro ↓5 66%
- 9
Gemma 3 27B ↓3 63.6%
- 10
Gemma 3 12B ↓2 63.3%
IndicIFEval-Ground gu
16 models we track have a score here.
- 1 GPT-5 77.5%
- 2 GPT-5 mini 74.5%
- 3
Qwen3-14B ↑2 71.3%
- 4
Qwen3-8B ↑5 65.2%
- 5
Gemma 3 27B ↑1 64.7%
- 6
Qwen3-32B ↑1 64.2%
- 7
Gemma 3 12B ↑1 63.4%
- 8
Qwen3-4B ↑3 61.3%
- 9
Gemini 3 Pro ↓6 61%
- 10
Gemma 3 4B 60.7%
IndicIFEval-Ground hi
16 models we track have a score here.
- 1 GPT-5 77.9%
- 2 GPT-5 mini 75.5%
- 3
Qwen3-14B ↑2 73.1%
- 4
Qwen3-8B ↑5 64.3%
- 5
Gemma 3 4B ↑5 64.1%
- 6
Qwen3-4B ↑5 63.7%
- 7
Gemma 3 12B ↑1 63.5%
- 8
Qwen3-32B ↓1 62.4%
- 9
Gemini 3 Flash ↓5 61.9%
- 10
Gemma 3 27B ↓4 60%
IndicIFEval-Ground kn
16 models we track have a score here.
- 1 GPT-5 73.6%
- 2 GPT-5 mini 71.5%
- 3
Qwen3-14B ↑2 69.2%
- 4
Qwen3-8B ↑5 67.5%
- 5
Gemma 3 12B ↑3 63%
- 6
Qwen3-32B ↑1 61.3%
- 7
Gemini 3 Pro ↓4 61%
- 8
Gemma 3 27B ↓2 59.2%
- 9
Qwen3-4B ↑2 58.1%
- 10
Gemini 3 Flash ↓6 58.1%
IndicIFEval-Ground ml
16 models we track have a score here.
- 1 GPT-5 70.6%
- 2
Qwen3-14B ↑3 70.3%
- 3 GPT-5 mini ↓1 68.2%
- 4
Gemma 3 27B ↑2 63%
- 5
Qwen3-32B ↑2 61.5%
- 6
Qwen3-8B ↑3 60.9%
- 7
Gemma 3 12B ↑1 60.7%
- 8
Gemini 3 Pro ↓5 59.4%
- 9
Gemma 3 4B ↑1 57.3%
- 10
Gemini 3 Flash ↓6 56.2%
IndicIFEval-Ground mr
16 models we track have a score here.
- 1 GPT-5 mini ↑1 81.7%
- 2 GPT-5 ↓1 78%
- 3
Qwen3-14B ↑2 69.7%
- 4
Qwen3-32B ↑3 66.6%
- 5
Gemma 3 27B ↑1 66%
- 6
Qwen3-8B ↑3 65.4%
- 7
Gemma 3 4B ↑3 64%
- 8
Gemini 3 Flash ↓4 62.9%
- 9
Qwen3-1.7B ↑3 62.6%
- 10
Qwen3-4B ↑1 62.6%
IndicIFEval-Ground ne
16 models we track have a score here.
- 1
Qwen3-14B ↑4 75.7%
- 2 GPT-5 mini 75.7%
- 3
Qwen3-8B ↑6 73%
- 4 GPT-5 ↓3 72.4%
- 5
Qwen3-32B ↑2 70.4%
- 6
Gemma 3 12B ↑2 66.9%
- 7
Gemini 3 Flash ↓3 66.9%
- 8
Gemma 3 27B ↓2 65.1%
- 9
Gemini 3 Pro ↓6 64.2%
- 10
Gemma 3 4B 63.3%
IndicIFEval-Ground or
16 models we track have a score here.
- 1 GPT-5 80.4%
- 2 GPT-5 mini 78.8%
- 3
Gemma 3 27B ↑3 69.3%
- 4
Qwen3-32B ↑3 69.3%
- 5
Qwen3-8B ↑4 68.2%
- 6
Gemini 3 Pro ↓3 67.9%
- 7
Qwen3-14B ↓2 67.4%
- 8
Gemini 3 Flash ↓4 62.8%
- 9
Qwen3-4B ↑2 62.2%
- 10
Gemma 3 12B ↓2 62%
IndicIFEval-Ground pa
16 models we track have a score here.
- 1 GPT-5 78.5%
- 2 GPT-5 mini 77%
- 3
Qwen3-14B ↑2 67.5%
- 4
Gemma 3 27B ↑2 65.8%
- 5
Gemini 3 Pro ↓2 65.5%
- 6
Qwen3-32B ↑1 64.8%
- 7
Gemini 3 Flash ↓3 64.5%
- 8
Qwen3-8B ↑1 63.3%
- 9
Gemma 3 12B ↓1 62.1%
- 10
Gemma 3 4B 59.7%
IndicIFEval-Ground sa
16 models we track have a score here.
- 1 GPT-5 mini ↑1 79.5%
- 2 GPT-5 ↓1 78.3%
- 3
Qwen3-4B ↑8 67%
- 4
Gemma 3 4B ↑6 65.8%
- 5
Qwen3-14B 64.3%
- 6
Qwen3-8B ↑3 62.8%
- 7
Qwen3-32B 61.3%
- 8
Gemma 3 27B ↓2 59.2%
- 9
Gemini 3 Flash ↓5 55.6%
- 10
Gemma 3 12B ↓2 54.2%
IndicIFEval-Ground ta
16 models we track have a score here.
- 1 GPT-5 mini ↑1 79.1%
- 2 GPT-5 ↓1 78.1%
- 3
Qwen3-14B ↑2 76.7%
- 4
Qwen3-8B ↑5 71.2%
- 5
Qwen3-32B ↑2 70%
- 6
Gemma 3 12B ↑2 67.9%
- 7
Gemini 3 Pro ↓4 66%
- 8
Gemini 3 Flash ↓4 64.9%
- 9
Qwen3-4B ↑2 61.2%
- 10
Gemma 3 27B ↓4 60.7%
IndicIFEval-Ground te
16 models we track have a score here.
- 1 GPT-5 mini ↑1 78.7%
- 2 GPT-5 ↓1 78.4%
- 3
Qwen3-32B ↑4 71.2%
- 4
Gemini 3 Pro ↓1 70.4%
- 5
Qwen3-14B 68.5%
- 6
Gemma 3 12B ↑2 68%
- 7
Qwen3-8B ↑2 67.2%
- 8
Gemma 3 27B ↓2 66.4%
- 9
Qwen3-4B ↑2 64.8%
- 10
Gemini 3 Flash ↓6 64.3%
IndicIFEval-Ground ur
16 models we track have a score here.
- 1 GPT-5 mini ↑1 73.6%
- 2 GPT-5 ↓1 72.7%
- 3
Qwen3-14B ↑2 60.2%
- 4
Qwen3-8B ↑5 60.2%
- 5
Gemini 3 Pro ↓2 60.2%
- 6
Qwen3-4B ↑5 59.9%
- 7
Gemini 3 Flash ↓3 59.6%
- 8
Qwen3-32B ↓1 57.6%
- 9
Gemma 3 27B ↓3 54.3%
- 10
Gemma 3 4B 54%
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: The IndicIFEval paper
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official IndicIFEval 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
- 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 11 months ago, in December 2025, for Gemini 3 Flash. 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 (19 entries).
- Pass Models scored 16 models we track, plus 3 other entries on its leaderboard.
- Warning Newest score The newest came 11 months ago, in December 2025, for Gemini 3 Flash.
- Note Frontier models None of the top 10 has a score. The strongest tested is Gemini 3 Pro (#56).
- Pass Headroom The best is 82.4%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on arxiv.org: 16 rows for models we track and 3 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.