Cohere Embed 5 launch: ViDoRe V3 parsed-text average

We show this table for reference; we do not rank on it.

Cohere launch retrieval report.

Evaluation results published by Cohere.

Should you trust a Cohere Embed 5 launch: ViDoRe V3 parsed-text average score?

Read with care: scores cluster at the top; the benchmark is not marked current.

Saturation Active: Top three within 2.1 score units; frontier rows are clustering, so read gaps at the top with care.
Who produced the rows Mirrored table: Every row is mirrored from Cohere Embed 5 launch report as it publishes it; BenchLM did not re-run these results. Source.
Freshness and weight Display only: Dated provider configurations, refreshed reviewed dated source capture. Display only; it does not move rankings. Upstream datasets and terms; aggregate results only.

Row statuses come from the verification registry; saturation and freshness from the benchmark lifecycle. A provider self-report is displayable evidence, not an independent run.

RCP-nDCG@10 on Cohere Embed 5 launch: ViDoRe V3 parsed-text average — September 30, 2026 launch; October 1 capture

We mirror the published rcp-ndcg@10 view for Cohere Embed 5 launch: ViDoRe V3 parsed-text average. Cohere Embed 5 Pro leads the public snapshot at 85.8, followed by Cohere Embed 5 Fast (84.5) and Voyage 4 Large (83.7). We do not use these results to rank models overall.

Cohere Embed 5 Pro

Cohere

Cohere provider report; evaluated snapshot and hardware undisclosed

85.8

Cohere Embed 5 Fast

Cohere

Cohere provider report; evaluated snapshot and hardware undisclosed

84.5

Voyage 4 Large

Voyage AI

Cohere provider report; evaluated snapshot and hardware undisclosed

83.7

11 configurationsCross-domainDated provider configurationsDisplay onlyUpdated September 30, 2026 launch; October 1 capture

RCP-nDCG@10 table (11 configurations)

RCP-nDCG@10 results for Cohere Embed 5 launch: ViDoRe V3 parsed-text average
Rank	Model / configuration	Score	Parameters (B)	Open / closed
1	Cohere Embed 5 ProCohereCohere provider report; evaluated snapshot and hardware undisclosed	85.8	Not reported	Unknown
2	Cohere Embed 5 FastCohereCohere provider report; evaluated snapshot and hardware undisclosed	84.5	Not reported	Unknown
3	Voyage 4 LargeVoyage AICohere provider report; evaluated snapshot and hardware undisclosed	83.7	Not reported	Unknown
4	Gemini Embedding 2GoogleCohere provider report; evaluated snapshot and hardware undisclosed	83.2	Not reported	Unknown
5	Voyage 4 NanoVoyage AICohere provider report; evaluated snapshot and hardware undisclosed	77.6	Not reported	Unknown
6	Cohere Embed 4CohereCohere provider report; evaluated snapshot and hardware undisclosed	77.0	Not reported	Unknown
7	OpenAI text-embedding-3-largeOpenAICohere provider report; evaluated snapshot and hardware undisclosed	75.5	Not reported	Unknown
8	Jina Embeddings v5 Text SmallJina AICohere provider report; evaluated snapshot and hardware undisclosed	74.5	Not reported	Unknown
9	Perplexity pplx-embed-v1PerplexityCohere provider report; evaluated snapshot and hardware undisclosed	74.4	Not reported	Unknown
10	Microsoft Harrier OSS v1MicrosoftCohere provider report; evaluated snapshot and hardware undisclosed	73.4	Not reported	Unknown
11	Qwen3-VL-EmbeddingQwenCohere provider report; evaluated snapshot and hardware undisclosed	64.2	Not reported	Unknown

How to read this provider report

Eight document domains, using parsed text curated by the ViDoRe authors. Values come from the two launch charts. RCP scores measure reordering a fixed candidate set; they do not measure first-stage retrieval recall.

Cohere reports these results in its Embed 5 launch. The linked full spreadsheet requires sign-in at capture, so this view includes only exact figures available in the public post. Model snapshots, evaluation dates, and uncertainty intervals are undisclosed.

Each input representation and metric retains its own cohort. RCP-nDCG@10 evaluates reordering a fixed candidate set; it does not measure first-stage retrieval. These provider comparisons do not enter weighted model rankings.

Snapshot

Cohere launch and charts Embed API model details Provider full-results sheet (sign-in required at capture)RCP-nDCG methodology and annotations

The published Cohere Embed 5 launch: ViDoRe V3 parsed-text average snapshot places Cohere Embed 5 Pro first at 85.8. The third row is 2.1 score units behind. The broader top-10 range is 12.4 score units, so the table still separates the published systems.

11 configurations are shown for Cohere Embed 5 launch: ViDoRe V3 parsed-text average. The benchmark falls in the Cross-domain category. We keep external benchmark mirrors separate from the weighted global scoring system, so these results remain source-specific evidence. Cohere Embed 5 launch: ViDoRe V3 parsed-text average is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About Cohere Embed 5 launch: ViDoRe V3 parsed-text average

Year

2026

Public benchmark source

Freshness and provenance

Version

Cohere Embed 5 launch provider report

Refresh cadence

Reviewed dated source capture

Staleness state

Dated provider configurations

Question availability

Upstream datasets and terms; aggregate results only

Dated provider configurationsDisplay only

BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.

Questions

What does Cohere Embed 5 launch: ViDoRe V3 parsed-text average measure?

Cohere launch retrieval report.

Which model leads the published Cohere Embed 5 launch: ViDoRe V3 parsed-text average snapshot?

Cohere Embed 5 Pro currently leads the published Cohere Embed 5 launch: ViDoRe V3 parsed-text average snapshot with 85.8 rcp-ndcg@10. BenchLM shows this benchmark for display only and does not use it in overall rankings.

How many models are evaluated on Cohere Embed 5 launch: ViDoRe V3 parsed-text average?

The September 30, 2026 launch; October 1 capture snapshot contains 11 AI models.

Last updated: September 30, 2026 launch; October 1 capture · mirrored from the public benchmark leaderboard

Know when it’s worth switching models

The model to choose, the cheaper alternative, and the release we would wait on.

Read a sample issue

Join 5,600+ readers.

One email each week. Unsubscribe anytime.