← All models

GreenPT green-embedding (Qwen3-Embedding-4B)

GreenPT

sustainable_balanced

Capabilities

Long ContextMultilingual

Strengths

  • 100% renewable energy inference
  • Heat recovery — server heat heats buildings
  • EU-hosted GDPR compliant
  • 100+ languages
  • Matryoshka — 32 to 2560 dim
  • Qwen3-Embedding-4B backbone (open model under green infrastructure)

Weaknesses

  • Subscription required for API
  • Single provider — smaller community than Cohere / OpenAI

Pricing

Input / 1M tokens

$0.22

Output

Hosting

Hosted only

Embedding specs

Output dimension

2560

Max input

32,000 tokens

Matryoshka

Supported

Matryoshka representation learning: dimension can be truncated to a smaller size at index time without retraining, trading retrieval quality for index size and speed.

Transparency

Open weights

10.0 / 10

Open training data

4.0 / 10

Open methodology

6.0 / 10

Licence openness

8.0 / 10

Provider disclosure

8.5 / 10

FMTI company score

N/A

Composite:7.5 / 10

Open-source backbone; hosting infrastructure documented.

Sustainability

Inference energy

9.8 / 10

Training footprint

N/A

Provider infrastructure

9.5 / 10

Composite:9.5 / 10

100% renewable energy + heat recovery. Best-in-class for hosted embedding.

MTEB quality

Embedding
80%

Grounded in the MTEB (Massive Text Embedding Benchmark) Overall average published by the model authors. Bearing collapses MTEB's retrieval, STS, classification, and clustering categories into a single quality signal because they correlate strongly for embedding models. Methodology.