Multilingual AI tools: compare listings

All Multilingual tools (23)

All Tags

Items tagged with Multilingual (23)

MiniMax M2.1

Open-source 230B parameter MoE model optimized for multi-language coding, agentic workflows, and real-world development tasks with 74% SWE-bench performance.

#Open Source

500000DR: 68AS: 72

EmbeddingGemma

Models

Lightweight multilingual text embedding model from Google DeepMind, optimized for on-device AI with <200MB RAM usage.

#Google

100000DR: 95AS: 95

Qwen3-Embedding

Models

Previous Qwen3 embedding snapshot.

#Alibaba

50000DR: 85AS: 88

Whisper V3

Models

Open-weight Whisper large-v3 family. Not OpenAI's current hosted STT default.

#OpenAI

8000DR: 95AS: 92

Cohere Embed v3

Models

Previous Cohere embed snapshot. Not a $0.10/1M 2026 default.

#Cohere

4500DR: 88AS: 86

Cohere Rerank 3.5

Models

Previous Cohere rerank snapshot.

#Cohere

4000DR: 88AS: 86

DeepSeek-Coder-V2.5

Models

Previous DeepSeek coder snapshot. Not a 90.2% HumanEval census.

#DeepSeek

3500DR: 88AS: 90

DeepSeek V2.5

Models

Previous DeepSeek V2.5 snapshot. Not a 2026 chat+coder default.

#DeepSeek

100DR: 100AS: 100

rerank-multilingual-v3.0

Models

Previous Cohere multilingual rerank snapshot.

#Cohere

238DR: 93AS: 88

Cohere: rerank-v3.5

Models

Previous Cohere rerank-v3.5 snapshot.

#Cohere

230DR: 93AS: 88

Jina AI Reranker v3

Models

Previous Jina rerank snapshot.

#Jina AI

450DR: 85AS: 82

BGE-M3

Models

2024 BAAI multilingual embedding snapshot. Not a 2026 MIRACL or MTEB crown.

#BAAI

—DR: 85AS: 80

Breeze TTS 2

Models

Open-weight multilingual TTS from BreezeBlue, built for real-time interaction. Reference-free voice design and reference-guided voice direction, with sub-40m.

#Text-to-Speech

—DR: —AS: —

Cohere Command A+

Models

Command A+ is Cohere's first mixture-of-experts flagship: 218B total and 25B active parameters, 48 languages, vision input, Apache 2.0 open weights.

#Cohere

—DR: —AS: —

Google: Gemini 3.5 Live Translate

Models

Google Gemini 3.5 Live Translate: leftover live-translate sibling on the live Gemini models page. Not the 2026 Flash default. No invent.

#Google

—DR: —AS: —

Google: Gemini 3.8 Flash TTS

Models

Gemini 3.8 Flash TTS designs custom voices from a prompt, replicates a voice from 30 seconds of audio, and leads Hume's voice design benchmark at 71.4.

#Google

—DR: —AS: —

Index-Translate

Models

Bilibili's open multilingual translation family: 150-language text, speech subtitles, controlled dubbing, and long-document translation built on Qwen3.5.

#Multilingual

—DR: —AS: —

Liquid d1-3B

Models

Liquid AI's 3B multimodal decision model for calibrated classification and edge-oriented image-text workflows.

#Open Source

—DR: —AS: —

Orukeet

Models

Orukeet is a 25-language ASR model that freezes fitted Gabor kernels on NVIDIA Parakeet TDT 0.6B v3.

#Speech-to-Text

—DR: —AS: —

Qwen-Audio-3.1

Models

Alibaba's Qwen-Audio-3.1 stack adds ASR-Next and TTS-Next, with TTS topping the Artificial Analysis TTS leaderboard and API prices cut up to 95%.

#Qwen

—DR: —AS: —

sanoTTS

Models

Tiny neural TTS from 294k to 2.3M parameters. Runs in the browser or on a $3 ESP32-S3. 11 voices, 6 languages, GPL-3.0.

#Text-to-Speech

—DR: —AS: —

Upstage: Solar Mini 4

Models

Upstage's 35B mixture-of-experts model with 3B active parameters, a 512K context window, and $0.10 per million input tokens.

#API

—DR: —AS: —

Spark-X2.5-4B

Models

XHToken's efficient 4B open-weight model: hybrid sliding-window attention for native 1M-token context, strong coding and agent performance, Apache 2.0, multilingual.

#Open Source

—DR: —AS: —