BRIGHT-Pro
Models retrieve passages that cover different parts of a question. The 739 questions include 115 from Earth Science Stack Exchange.
MTEB ranks 19 embedding models and a keyword baseline across seven domains. The authors also test agents that search repeatedly before answering.
- Made by
- Yale NLP and collaborators
- Released
- April 2026
- Kind
- Questions that need search
- Level
- University
- Scored
- Checked automatically
- Answers
- Short answer
- Tasks
- All public
- Measures
- Capability
- Subject
- Earth and environmental sciences
- Language
- English
- Interaction
- Agentic
- Environment
- APIs and tools
- Items from
- Real work and use
- Refresh
- One fixed set
- Run by
- Anyone
- Based on
- BRIGHT
- In views
- Agents
- Last model added
- June 2025
- Scores refreshed
- November 20th, 2026
AQ-MedAI/Diver-Retriever-4B-1020 tops the leaderboard with 55.9 Qwen3-Embedding-8B leads with 41.8 among models released by June 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. The best model we track is Qwen3-Embedding-8B, at 41.8. This view shows only models we track, each at its best.
The newest model with a score here came out in June 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
- 55.9 AQ-MedAI/Diver-Retriever-4B-1020
- 52.0 AQ-MedAI/Diver-Retriever-4B
- 49.0 AQ-MedAI/Diver-Retriever-0.6B
- 48.0 AQ-MedAI/Diver-Retriever-1.7B
- 46.9 yale-nlp/RTriever-4B
- 41.8
Qwen3-Embedding-8B
- 40.8
Qwen3-Embedding-4B
- 38.3 DataScience-UIBK/uibk-embedding-v1
- 37.8 Alibaba-NLP/gte-Qwen2-7B-instruct
- 32.9 ReasonIR/ReasonIR-8B
- 29.9
Qwen3-Embedding-0.6B
- 29.8 BAAI/bge-large-en-v1.5
- 28.6 sentence-transformers/all-mpnet-base-v2
- 24.9 intfloat/multilingual-e5-large
- 24.6 codefuse-ai/F2LLM-v2-80M
- 23.5 mteb/baseline-bm25s
- 21.1 minishlab/potion-multilingual-128M
- 20.5 intfloat/multilingual-e5-small
- 20.5 BAAI/bge-m3
- 19.4 sentence-transformers/static-similarity-mrl-multilingual-v1
- 1 AQ-MedAI/Diver-Retriever-4B-1020 55.9
- 2 AQ-MedAI/Diver-Retriever-4B 52.0
- 3 AQ-MedAI/Diver-Retriever-0.6B 49.0
- 4 AQ-MedAI/Diver-Retriever-1.7B 48.0
- 5 yale-nlp/RTriever-4B 46.9
- 6
Qwen3-Embedding-8B 41.8
- 7
Qwen3-Embedding-4B 40.8
- 8 DataScience-UIBK/uibk-embedding-v1 38.3
- 9 Alibaba-NLP/gte-Qwen2-7B-instruct 37.8
- 10 ReasonIR/ReasonIR-8B 32.9
- 11
Qwen3-Embedding-0.6B 29.9
- 12 BAAI/bge-large-en-v1.5 29.8
- 13 sentence-transformers/all-mpnet-base-v2 28.6
- 14 intfloat/multilingual-e5-large 24.9
- 15 codefuse-ai/F2LLM-v2-80M 24.6
- 16 mteb/baseline-bm25s 23.5
- 17 minishlab/potion-multilingual-128M 21.1
- 18 intfloat/multilingual-e5-small 20.5
- 19 BAAI/bge-m3 20.5
- 20 sentence-transformers/static-similarity-mrl-multilingual-v1 19.4
Source: Official BRIGHT-Pro leaderboard
- 1
Qwen3-Embedding-8B 41.8
- 2
Qwen3-Embedding-4B 40.8
- 3
Qwen3-Embedding-0.6B 29.9
Source: Official BRIGHT-Pro leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | AQ-MedAI/Diver-Retriever-4B-1020 open weights, 4B | 55.9 | Official leaderboard | ||
| 2 | AQ-MedAI/Diver-Retriever-4B open weights, 4B | 52.0 | Official leaderboard | ||
| 3 | AQ-MedAI/Diver-Retriever-0.6B open weights, 0.6B | 49.0 | Official leaderboard | ||
| 4 | AQ-MedAI/Diver-Retriever-1.7B open weights, 1.7B | 48.0 | Official leaderboard | ||
| 5 | yale-nlp/RTriever-4B open weights, 4B | 46.9 | Official leaderboard | ||
| 6 | | 41.8 | Jun 2025 | Official leaderboard | |
| 7 | | 40.8 | Jun 2025 | Official leaderboard | |
| 8 | DataScience-UIBK/uibk-embedding-v1 open weights, 0.2B | 38.3 | Official leaderboard | ||
| 9 | Alibaba-NLP/gte-Qwen2-7B-instruct open weights, 7.1B | 37.8 | Official leaderboard | ||
| 10 | ReasonIR/ReasonIR-8B open weights, 7.5B | 32.9 | Official leaderboard | ||
| 11 | | 29.9 | Jun 2025 | Official leaderboard | |
| 12 | BAAI/bge-large-en-v1.5 open weights, 0.3B | 29.8 | Official leaderboard | ||
| 13 | sentence-transformers/all-mpnet-base-v2 open weights, 0.1B | 28.6 | Official leaderboard | ||
| 14 | intfloat/multilingual-e5-large open weights, 0.6B | 24.9 | Official leaderboard | ||
| 15 | codefuse-ai/F2LLM-v2-80M open weights, 0.1B | 24.6 | Official leaderboard | ||
| 16 | mteb/baseline-bm25s open weights | 23.5 | Official leaderboard | ||
| 17 | minishlab/potion-multilingual-128M open weights, 0.1B | 21.1 | Official leaderboard | ||
| 18 | intfloat/multilingual-e5-small open weights, 0.1B | 20.5 | Official leaderboard | ||
| 19 | BAAI/bge-m3 open weights, 0.6B | 20.5 | Official leaderboard | ||
| 20 | sentence-transformers/static-similarity-mrl-multilingual-v1 open weights, 0.1B | 19.4 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 41.8 | Jun 2025 | Official leaderboard | |
| 2 | | 40.8 | Jun 2025 | Official leaderboard | |
| 3 | | 29.9 | Jun 2025 | Official leaderboard |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official BRIGHT-Pro 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 BRIGHT-Pro 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 Knowledge
- MMLU-Pro In use Harder 10-option version of MMLU with reasoning-focused questions across 14 subjects.
Intern-S2-397B 89.8%
- MMLU In use 57-subject multiple-choice exam questions from elementary to professional level. Scored by accuracy.
Seed1.8 92.3%
- SuperGPQA In use 26,529 graduate-level multiple-choice questions across 285 disciplines. Accuracy.
Qwen 3.6 Max (Preview) 73.9%
- WideSearch In use 200 bilingual broad info-seeking tasks: agent collects many atomic facts into a table. Scored by success rate and row/item F1.
Tencent Hy4 preview 83.9%
- BBEH micro average In use Models solve harder replacements for each of the 23 BIG-Bench Hard tasks, 4,520 examples in total. The score is the share of all examples answered correctly.
Gemma 4 31B IT 74.4%
- HLE-Diamond In use Models answer 1,000 cleaned expert questions from Humanity's Last Exam, 500 on reasoning and 500 on knowledge, closed book. Graded on exact answers. Claude Fable 5.2 77.9%
Data audit
8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 17 months ago, in June 2025, for Qwen3-Embedding-0.6B. 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 (20 entries).
- Pass Models scored 3 models we track, plus 17 other entries on its leaderboard.
- Warning Newest score The newest came 17 months ago, in June 2025, for Qwen3-Embedding-0.6B.
- Note Frontier models None of its models are on the capabilities index.
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on leaderboard.mteb.org: 3 rows for models we track and 17 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.