Other
whisper/tokenizer.py: the LANGUAGES table (openai/whisper)
OpenAI, GitHub, openai/whisper, 2026
Read the originalRead on October 16, 2026
- What it is
- The source file of OpenAI's Whisper speech recognition package that holds its LANGUAGES table, the list of language codes the model accepts, such as bn for Bengali, ur for Urdu and ar for Arabic.
- How much weight it carries
- The project's own source code, the authoritative list for the open Whisper package. It shows which languages the model was built to handle, not how accurate it is in each one. Apps built on Whisper may enable only some of these languages.
Facts taken from this source
- The LANGUAGES table in OpenAI's Whisper source code includes Arabic, with the code ar.
“"nl": "dutch", "ar": "arabic", "sv": "swedish", "it": "italian"” (LANGUAGES dictionary)
- The LANGUAGES table in OpenAI's Whisper source code includes Bengali, with the code bn.
“"lv": "latvian", "bn": "bengali", "sr": "serbian", "az": "azerbaijani"” (LANGUAGES dictionary)
- The LANGUAGES table in OpenAI's Whisper source code includes Bengali, with the code bn, and Urdu, with the code ur.
“"th": "thai", "ur": "urdu", "hr": "croatian", "bg": "bulgarian"... "lv": "latvian", "bn": "bengali", "sr": "serbian", "az": "azerbaijani"” (LANGUAGES dictionary)
- Whisper's language table has one code for German, de, and one for Portuguese, pt, with no separate Swiss German code and no separate Brazilian or European Portuguese code.
“"de": "german", "es": "spanish", "ru": "russian", "ko": "korean", "fr": "french", "ja": "japanese", "pt": "portuguese"” (LANGUAGES table)
- The language table in Whisper's source code includes Shona, Yoruba, Somali and Afrikaans, each with its own language code.
“"sn": "shona", "yo": "yoruba", "so": "somali", "af": "afrikaans",” (LANGUAGES dictionary in tokenizer.py)
- The language table in Whisper's source code includes Chinese, Korean and Japanese, each with its own language code.
“"zh": "chinese", "de": "german", "es": "spanish", "ru": "russian", "ko": "korean", "fr": "french", "ja": "japanese",” (LANGUAGES dictionary in tokenizer.py)
- The LANGUAGES table in OpenAI's Whisper source code includes Urdu, with the code ur.
“"th": "thai", "ur": "urdu", "hr": "croatian", "bg": "bulgarian"” (LANGUAGES dictionary)
Pages that rest on this source
- Arabic Speech to Text and Right-to-Left DictationGuide
- Can you voice type in Bangla, Urdu or Arabic on Windows?Question
- Does Whisper support Arabic?Question
- What is the best Arabic speech to text API?Question
- How do you voice type in Bangla and get English text?Question
- Speech to Text in French, German, Russian and MoreGuide
- Does dictation work for both Brazilian and European Portuguese?Question
- Does speech recognition understand Swiss German?Question
- Can AI transcribe Zulu, Xhosa and other African languages?Question
- Japanese, Chinese and Korean Speech to TextGuide
- How do you transcribe Japanese, Chinese or Korean audio?Question
- How do you voice type in Urdu and get English text?Question
- How do you voice type in Urdu?Question