Chinese generative AI models (DeepSeek and Qwen) rival ChatGPT-4 in ophthalmology queries with excellent performance in Arabic and English
Bibliographic record
Abstract
The rapid evolution of generative artificial intelligence (genAI) has ushered in a new era of digital medical consultations, with patients turning to AI-driven tools for guidance. The emergence of Chinese-developed genAI models such as DeepSeek-R1 and Qwen-2.5 presented a challenge to the dominance of OpenAI’s ChatGPT. The aim of this study was to benchmark the performance of Chinese genAI models against ChatGPT-4o and to assess disparities in performance across English and Arabic. Following the METRICS checklist for genAI evaluation, Qwen-2.5, DeepSeek-R1, and ChatGPT-4o were assessed for completeness, accuracy, and relevance using the CLEAR tool in common patient ophthalmology queries. In English, Qwen-2.5 demonstrated the highest overall performance (CLEAR score: 4.43±0.28), outperforming both DeepSeek-R1 (4.31±0.43) and ChatGPT-4o (4.14±0.41), with p=0.002. A similar hierarchy emerged in Arabic, with Qwen-2.5 again leading (4.40±0.29), followed by DeepSeek-R1 (4.20±0.49) and ChatGPT-4o (4.14±0.41), with p=0.007. Each tested genAI model exhibited near-identical performance across the two languages, with ChatGPT-4o demonstrating the most balanced linguistic capabilities (p=0.957), while Qwen-2.5 and DeepSeek-R1 showed a marginal superiority for English. An in-depth examination of genAI performance across key CLEAR components revealed that Qwen-2.5 consistently excelled in content completeness, factual accuracy, and relevance in both English and Arabic, setting a new benchmark for genAI in medical inquiries. Despite minor linguistic disparities, all three models exhibited robust multilingual capabilities, challenging the long-held assumption that genAI is inherently biased toward English. These findings highlight the evolving nature of AI-driven medical assistance, with Chinese genAI models being able to rival or even surpass ChatGPT-4o in ophthalmology-related queries.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".