Cohere says Embed 5 Fast beats Google's EmbeddingGemma 2 on 95% of 163 benchmark datasets
cohere · x · 2026-10-18
Cohere benchmarked its Embed 5 Fast against Google's EmbeddingGemma 2 across 163 datasets using RCP-nDCG@10, claiming leads on over 95% of them. Charts show ViDoRe v3 RCP average of 84.5 vs 71.4, and 81.2 vs 71.7 on the main text-search benchmark.
- Responding to questions about RCP-nDCG being its own methodology, Cohere says it still leads with standard nDCG@10 (ViDoRe v3 Regular: 59.3 vs 47.6).
- Vendor self-benchmark with self-chosen metrics; independent validation still pending.
Related event: Cohere Claims Embed 5 Fast Beats Google's Embedding Model Across Benchmarks(2 posts)→
More from Models
- DeepSeek V4.1 Flash as a coding agent: fast at the core, but prone to rabbit-holing on polish — teortaxesTex · 2026-10-18
- Users report Codex usage limits draining abnormally fast, 16% gone in seconds — Angaisb_ · 2026-10-18
- OpenAI's Codex weekly usage reset rolls out, limits now shown as 100% left — ChrisUniverse · 2026-10-18
- TeortaxesTex: DeepSeek best positioned to recover from distillation, but Claude personas will linger — teortaxesTex · 2026-10-18
- Elorian's Glimpse matches Argon on labels but costs 25x more than Gemma 4 31B — arnabdotorg · 2026-10-18
- 5-model blind test: DeepSeek builds Adidas shoe site for $0.23 vs Opus 5.5's $33 — cedric_chee · 2026-10-18