IndicContextEval
Models transcribe speech in eight Indian languages using hints about the topic or names. One prompt deliberately supplies names from an unrelated field.
The authors publish scores for five models with a language hint. Four models also receive the other six prompts on the same speech.
- Made by
- AI4Bharat at IIT Madras and Sarvam AI
- Released
- June 2026
- Kind
- Speech recognition and translation
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Open-ended
- Tasks
- All public
- Measures
- Capability
- Subject
- Languages and literature
- Language
- Many languages
- Interaction
- Single turn
- Items from
- Real work and use, Written by experts
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Audio and speech, Many languages
- Strongest model tested
-
Gemini 3 Flash #60 on the capability index
- Last model added
- December 2025
- Scores refreshed
- November 20th, 2026
Sarvam Audio tops the leaderboard with 15.7% Gemini 3 Flash leads with 17.5% among models released by December 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. Every run is in the next chart. 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 Gemini 3 Flash, at 17.5%. 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
- 1 Sarvam Audio L5: Language + entities (native script) 15.7%
- 2
Gemini 3 Flash L5: Language + entities (native script) 17.5%
- 3 IndicConformer L1: Language only (baseline) 18.8%
- 4 GPT-4o Mar 2025, L5: Language + entities (native script) 26%
- 5
Gemma 3n E4B L1: Language only (baseline) 38.7%
Source: Official IndicContextEval leaderboard
Source: Official IndicContextEval leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Sarvam Audio | 15.7% | L5: Language + entities (native script) | Official leaderboard | |
| 2 | | 17.5% | L5: Language + entities (native script) | Dec 2025 | Official leaderboard |
| 3 | IndicConformer | 18.8% | L1: Language only (baseline) | Official leaderboard | |
| 4 | GPT-4o Mar 2025 | 26% | L5: Language + entities (native script) | Mar 2025 | Official leaderboard |
| 5 | | 38.7% | L1: Language only (baseline) | Jun 2025 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 17.5% | L5: Language + entities (native script) | Dec 2025 | Official leaderboard |
| 2 | GPT-4o Mar 2025 | 26% | L5: Language + entities (native script) | Mar 2025 | Official leaderboard |
| 3 | | 38.7% | L1: Language only (baseline) | Jun 2025 | Official leaderboard |
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: Official IndicContextEval leaderboard
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| L2: Language + domain metadata | 52.2% | ||||
| L0: No context | 51.2% | ||||
| L6: Wrong entities (adversarial) | 48% | ||||
| L4: Language + entities (English) | 46.4% | ||||
| L5: Language + entities (native script) | 43.1% | ||||
| L3: Language + audio description | 40.2% | ||||
| L1: Language only (baseline) | 38.7% | ||||
| GPT-4o Mar 2025 | L0: No context | 29.8% | |||
| GPT-4o Mar 2025 | L1: Language only (baseline) | 28.6% | |||
| GPT-4o Mar 2025 | L6: Wrong entities (adversarial) | 28.5% | |||
| GPT-4o Mar 2025 | L2: Language + domain metadata | 28.4% | |||
| GPT-4o Mar 2025 | L4: Language + entities (English) | 28% | |||
| GPT-4o Mar 2025 | L3: Language + audio description | 26.1% | |||
| GPT-4o Mar 2025 | L5: Language + entities (native script) | 26% | |||
| L0: No context | 24.3% | ||||
| L4: Language + entities (English) | 19.9% | ||||
| L6: Wrong entities (adversarial) | 19.7% | ||||
| L2: Language + domain metadata | 19.3% | ||||
| L1: Language only (baseline) | 18.9% | ||||
| L3: Language + audio description | 18.4% | ||||
| L5: Language + entities (native script) | 17.5% |
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: Official IndicContextEval leaderboard
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| L2: Language + domain metadata | 52.2% | ||||
| L0: No context | 51.2% | ||||
| L6: Wrong entities (adversarial) | 48% | ||||
| L4: Language + entities (English) | 46.4% | ||||
| L5: Language + entities (native script) | 43.1% | ||||
| L3: Language + audio description | 40.2% | ||||
| L1: Language only (baseline) | 38.7% | ||||
| GPT-4o Mar 2025 | L0: No context | 29.8% | |||
| GPT-4o Mar 2025 | L1: Language only (baseline) | 28.6% | |||
| GPT-4o Mar 2025 | L6: Wrong entities (adversarial) | 28.5% | |||
| GPT-4o Mar 2025 | L2: Language + domain metadata | 28.4% | |||
| GPT-4o Mar 2025 | L4: Language + entities (English) | 28% | |||
| GPT-4o Mar 2025 | L3: Language + audio description | 26.1% | |||
| GPT-4o Mar 2025 | L5: Language + entities (native script) | 26% | |||
| L0: No context | 24.3% | ||||
| L4: Language + entities (English) | 19.9% | ||||
| L6: Wrong entities (adversarial) | 19.7% | ||||
| L2: Language + domain metadata | 19.3% | ||||
| L1: Language only (baseline) | 18.9% | ||||
| L3: Language + audio description | 18.4% | ||||
| L5: Language + entities (native script) | 17.5% |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official IndicContextEval leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official IndicContextEval 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 Audio and video
- Video-MMMU In use 300 expert-level lecture videos with questions testing perception, comprehension and adaptation of knowledge. Accuracy.
Qwen 3.8 Max 88.7%
- LVBench In use Hours-long video understanding with multiple-choice questions on reasoning, grounding and summarization.
Gemini 4 Argon 91.7%
- MMVU In use 3,000 expert-annotated questions on videos across 27 disciplines. Multiple-choice and open-ended accuracy.
Qwen 3.8 Max 82.4%
- MLVU In use Multi-task long video understanding (minutes to hours) with multiple-choice and generation tasks. Accuracy.
Qwen 3.8 Max 90.8%
- FLEURS In use Models transcribe read speech in 102 languages built from FLoRes-101 sentences. Model cards usually report word error rate, so lower scores are better.
Phi-4-multimodal 4.00%
- VoiceBench In use Spoken instructions with speaker, environment and content variation. Voice assistant responses scored by GPT judge or rules.
Qwen3.5-Omni-Plus 93.1%
Data audit
8 of 9 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 2 pass 2 notes
- Pass Independent runs Scores from the official leaderboard (5 entries).
- Pass Models scored 3 models we track, plus 2 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 Flash (#60).
- Note Headroom Lower is better, so there’s no ceiling to reach.
Leaderboard import 3 pass
- Pass Importer We read the board on github.com: 21 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.