Qwen3-Embedding-4B
Alibaba Qwen · open weights
$0.020 per 1M input tokens, cheapest of 2 sellers (DeepInfra)
What was measured
| Board | Mean (Task) | Mean (TaskType) | Tasks | Revision |
|---|---|---|---|---|
| MTEB(eng, v2) | 74.61 | 68.1 | 41 of 41 | 636cd9bf47d976946cdbb2b0c3ca0cb2f8eea5ff |
By task type: Classification 89.84 · Clustering 57.51 · PairClassification 87.01 · Reranking 50.76 · Retrieval 68.46 · STS 88.72 · Summarization 34.39
What sellers charge
Spread across 2 distinct sellers: 1× between cheapest and dearest at the reference unit.
| Seller | As the seller states it | Conditions | Normalised per 1M input tokens | Source |
|---|---|---|---|---|
| DeepInfra | 2e-8 per_token | — | $0.020 per token × 1,000,000 | api.deepinfra.com read 2026-09-19 · feed |
| OpenRouter | 2e-8 per_token | context 32768 | $0.020 per token × 1,000,000 | openrouter.ai read 2026-09-19 · feed |
Provenance
- Quality: https://github.com/embeddings-benchmark/results, licensed cc0-1.0.
- Prices: https://api.deepinfra.com/models/list, https://openrouter.ai/api/v1/embeddings/models
- Built 2026-09-19 from the catalogue snapshot in this repository.
MTEB results, https://github.com/embeddings-benchmark/results, released under CC0-1.0; MTEB(eng, v2) means computed from the per-task files as described in data/modalities/embeddings/benchmark.json.