Skip to main content
Vibgrate Graph: over 27% fewer tokens in our benchmark
Back to Tags

Multilingual

4 items tagged with "multilingual"

Filter by type:

Models3

Model

Qwen3.8 Max Prime

Qwen3.8 Max Prime is an Alibaba Qwen hosted foundation model added to OpenRouter with a 1,000,000-token context window. It is a Prime variant of the Qwen3.8 Max line for long-context general reasoning and generation.

Model

DeepSeek V3

MoE model with 671B parameters achieving frontier performance

Model

Llama 3.3 70B

Open-weight multilingual model matching Llama 3.1 405B performance

Benchmarks1

Benchmark

MGSM (Multilingual Grade School Math)

A multilingual extension of grade-school math word problems that tests whether LLMs can reason through arithmetic in many languages, not just English.