Mixedbread Quality Evals
How Mixedbread performs on public retrieval and agent benchmarks, measured against the strongest competing models.
Every result comes from a public benchmark on its standard split and evaluation script. Retrieval numbers are Wholembed V3 served through the Stores API; agent numbers come from Toast 1 or third-party agents running on Mixedbread Search, with cost and latency measured per query at list prices. Each benchmark lists its dataset, exact methodology and testing dates.
How we measure
Every result comes from a public benchmark on its standard split and evaluation script. Retrieval numbers are Wholembed V3 served through the Stores API; agent numbers come from Toast 1 or third-party agents running on Mixedbread Search, with cost and latency measured per query at list prices. Each benchmark lists its dataset, exact methodology and testing dates.
Wholembed V3, served through the Stores API, against other embedding models on public retrieval evals.
Audio retrieval with graded relevance judgements.
Per-metric breakdown
| Model | Lenient Open Lenient dataset | Strict Open Strict dataset |
|---|---|---|
| Mixedbread | 77.3 | 65.1 |
| Gemini Embedding 2 | 72.5 | 64.0 |
| CLAMP3 | 70.4 | 57.9 |
| ColQwen-Omni | 65.4 | 51.4 |
| TTMR++ | 64.5 | 50.2 |
| CLAP | 59.4 | 46.0 |
NDCG@10 · higher is betterTested March 2026Incompebench (Lenient)Incompebench (Strict)
Dataset and methodology
Incompebench uses a 0–3 graded relevance scale. Strict NDCG@10 treats only ratings 2 and 3 as positive (tangentially-relevant tracks count as negative). Lenient NDCG@10 uses all three relevance grades, giving partial credit for somewhat-relevant matches.
We evaluate retrieval quality using publicly available eval datasets. Each model encodes the corpus and queries using its standard inference pipeline, then we compute the reported metric over the full test split. We report results as published by the eval or as measured on identical splits with default evaluation scripts.