This study explores the capabilities and limitations of artificial intelligence (AI) in literary translation by comparing translations of Halide Edib Adıvar's Sinekli Bakkal produced by ChatGPT, DeepL, and Gemini AI. ChatGPT and Gemini are open AI models with broad language capabilities, performing a wide range of tasks, not just translation but also text generation and language understanding, while DeepL is software specifically designed with a focus on translation. The study, in which thematic sampling method is utilized as a qualitative research method, it is aimed to highlight differences in achieving equivalence between source and target texts through comparative analysis of selected passages. It draws on Nida's concepts of formal and dynamic equivalence and Venuti's domestication and foreignization strategies to evaluate the effectiveness of AI tools. The human-versions of the target texts are not included in the study, since it focuses on the performance of different AI-assisted translation tools rather than comparing the human and AI in terms of translation abilities. The results indicate that while ChatGPT and Gemini AI balance readability and cultural nuance more effectively, DeepL often produces literal yet less nuanced translations. Despite improvements in AI translation, challenges remain in handling idiomatic expressions, humour, and cultural references, with errors such as gender mismatches underscoring the need for human intervention. The study concludes that AI tools can complement human translators but cannot replace the creativity, cultural sensitivity, and contextual understanding required for high-quality literary translations. Bu çalışma, Halide Edib Adıvar'ın Sinekli Bakkal adlı eserinin ChatGPT, DeepL ve Gemini AI tarafından yapılan çevirilerini karşılaştırarak yapay zekanın edebi çevirideki yeteneklerini ve sınırlamalarını araştırmaktadır. ChatGPT ve Gemini, geniş dil yeteneklerine sahip açık yapay zekâ modelleri olup, yalnızca çeviri değil, metin üretimi ve dil anlama gibi çok çeşitli görevleri de yerine getirirken, DeepL özel olarak çeviri odaklı tasarlanmış bir yazılımdır. Çalışmanın amacı, bu farklı türdeki yazılım araçlarının kaynak metindeki anlamı, duyguyu ve kültürel bağlamı hedef dile aktarımda nasıl bir performans sergilediğini incelemektir. Nitel bir araştırma yöntemi olan tematik örnekleme metodu kullanılan çalışmada, seçilen paragrafların karşılaştırmalı analizi yoluyla araştırma, kaynak ve hedef metinler arasında eşdeğerliğe ulaşmadaki farklılıkları vurgulamaktadır. Yapay zekâ araçlarının i Arş. Gör., Çağ Üniversitesi, Fen Edebiyat Fakültesi, İngilizce Mütercim ve Tercümanlık Bölümü, e-posta: bariscanaydin@cag.edu.tr, ORCID ID: 0000-0002-8268-6020. 82 etkinliğini değerlendirmek için Nida'nın biçimsel ve dinamik eşdeğerlik kavramlarından ve Venuti'nin yerlileştirme ve yabancılaştırma stratejilerinden yararlanmaktadır. Çalışma, insan ve yapay zekâ arasındaki çeviri yeteneğini karşılaştırmaktan ziyade, farklı yapay zeka destekli çeviri araçlarının performanslarını karşılaştırmayı amaçladığı için, tamamen insan tarafından çevrilmiş bir erek metin çalışmaya dahil edilmemiştir. Sonuçlar, ChatGPT ve Gemini AI'nın okunabilirlik ve kültürel farklılıkları daha etkili bir şekilde dengelerken, DeepL'nin genellikle tam anlamıyla ancak daha az nüanslı çeviriler ürettiğini göstermektedir. Yapay zekâ çevirisindeki gelişmelere rağmen, deyimsel ifadeleri, mizahı ve kültürel referansları ele almada zorluklar ve cinsiyet uyumsuzluğu gibi hatalar insan müdahalesine olan ihtiyacı vurgulamaktadır. Çalışma, yapay zekâ araçlarının insan çevirmenleri tamamlayabileceği ancak yüksek kaliteli edebi çeviriler için gereken yaratıcılığın, kültürel duyarlılığın ve bağlamsal anlayışın yerini alamayacağı sonucuna varmaktadır. Extended Abstract The advent of artificial intelligence (AI) has revolutionized the field of translation, enabling faster and more efficient processes. However, while AI-powered translation tools such as ChatGPT, DeepL, and Gemini AI have demonstrated remarkable progress in handling technical and informational texts, challenges remain in the field of literary translation. Literary texts often involve intricate stylistic elements, emotional undertones, and culturally specific references that require creativity and deep contextual understanding-qualities typically associated with human translators. This study investigates the effectiveness of AI-assisted translation in producing equivalent literary translations by examining multiple versions of Halide Edib Adıvar's Sinekli Bakkal, a prominent Turkish novel. This study compares ChatGPT, Gemini AI, and DeepL in the field of literary translation to investigate how AI tools with varying core functionalities perform when translating culturally rich and stylistically intricate texts. ChatGPT and Gemini are general-purpose AI language models with extensive linguistic capabilities, including translation, text generation, and contextual understanding. DeepL, however, is a specialized machine translation tool, widely recognized for its fluency and accuracy. The central research question guiding this investigation is: 'How do AI translation tools with different functionalities handle the challenges of literary translation, particularly in achieving equivalence and cultural adaptation?' The analysis involves selected passages from Halide Edib Adıvar's Sinekli Bakkal, focusing on idiomatic expressions, cultural references, and stylistic nuances. Drawing on Nida's formal and dynamic equivalence and Venuti's domestication and foreignization strategies, this study evaluates the ability of these tools to achieve equivalence while preserving the text's literary qualities. The findings reveal that while ChatGPT and Gemini balance readability and cultural adaptation effectively, DeepL excels in literal accuracy but struggles with nuanced cultural references. This highlights the complementary role AI can play alongside human translators in achieving high-quality literary translations. The research focuses on the concept of equivalence in literary translation, drawing on key theoretical frameworks, including Eugene Nida's formal and dynamic equivalence and Lawrence Venuti's strategies of domestication and foreignization. Through a qualitative comparative analysis, the study evaluates how different AI tools perform in balancing literal accuracy with readability, emotional resonance, and cultural fidelity. Specifically, selected passages from Sinekli Bakkal-rich with metaphors, idiomatic expressions, and character-driven narratives-were translated by ChatGPT, DeepL, and Gemini AI. These translations were compared to the original Turkish text, highlighting differences in style, tone, and semantic accuracy. The study examines how each tool manages the inherent challenges of literary translation, including idiomatic language, humor, gender references, and culturally specific imagery. The findings reveal both the strengths and limitations of AI-assisted translation in achieving equivalence. ChatGPT and Gemini AI demonstrate a greater ability to capture the dynamic aspects of the source text, focusing on meaning, narrative flow, and emotional impact. For instance, in character descriptions (e.g., Sabiha Hanım 83 and Tevfik), these tools deliver translations that convey both physical details and underlying emotional nuances. Their translations often strike a balance between literal fidelity and interpretative flexibility, making them more readable and engaging. In contrast, DeepL tends to prioritize formal equivalence, producing translations that, while structurally accurate, occasionally result in awkward phrasing or slight shifts in meaning. An example of this occurs in a passage where DeepL translates "imam's granddaughter" as "grandson," altering the narrative consistency. Such errors highlight AI's struggles with gender context and subtle shifts in meaning. Venuti's concepts of domestication and foreignization provide further insight into the stylistic differences between the translations. ChatGPT tends to employ a domesticated approach, rendering the text in ways that feel accessible to English-speaking readers while preserving key cultural references. In contrast, DeepL's literal translations align more with foreignization but occasionally diminish readability, as seen in the translation of phrases like "does not stop at one branch," which reads unnaturally in English. Gemini AI offers a middle ground, preserving many cultural elements while adapting the phrasing for clarity and fluency. This comparative analysis underscores the complexity of literary translation, where balancing the original text's cultural essence with the target audience's expectations requires nuanced decision-making-something AI tools are still refining. One of the most notable limitations identified in this study involves the treatment of idiomatic expressions, humour, and emotional undertones. AI-generated translations often struggle to capture humour and irony, as illustrated by a scene featuring Tevfik's humorous antics. While ChatGPT maintains the playful tone effectively, DeepL's literal approach leads to less fluid, slightly mechanical expressions. Furthermore, the analysis highlights that AI tools may misinterpret cultural metaphors, which human translators would instinctively adapt to the target language. This limitation supports the view that literary translation is not merely a linguistic exercise but also a cultural and emotional endeavour. Despite these limitations, the study recognizes the value of AI in enhancing translation efficiency. AI tools offer significant advantages in terms of speed and consistency, particularly for initial drafts or repetitive translation tasks. However, they still require human post-editing to correct errors, refine stylistic elements, and ensure the translation resonates with the target audience. This aligns with the concept of "human-in-the-loop" translation, where AI and human translators collaborate to produce high-quality outputs. Castilho et al. (2018) emphasize that this hybrid model is especially valuable in specialized fields like literary translation, where both linguistic precision and creativity are essential. The findings also suggest that AI models, while capable of handling a wide range of languages, perform unevenly across different linguistic contexts. In the case of Sinekli Bakkal, which reflects both Turkish and Ottoman cultural elements, the AI tools encountered challenges with idiomatic expressions and historical references. These limitations highlight the need for further research into improving AI's cultural sensitivity and contextual understanding, particularly for low-resource languages and texts with complex literary styles. In conclusion, this study demonstrates that AI tools like ChatGPT, DeepL, and Gemini AI have made significant progress in the field of literary translation but are not yet capable of fully replacing human translators. While they excel at generating readable and structurally sound translations, they struggle with idiomatic language, emotional depth, and cultural nuance. The results underscore the importance of human expertise in refining AI-generated texts, ensuring that the final translation captures both the letter and spirit of the original. Future research should focus on enhancing AI's contextual awareness and integrating feedback mechanisms between human translators and AI systems. Ultimately, AI will continue to complement human translators, fostering new possibilities for collaboration and innovation in literary translation.