DeepL published a new blog explaining why OpenAI GPT-5.2 is 4x slower than DeepL, Claude Opus 4.6 is 6x slower, and Google Gemini 3.1 Pro is a vast 29x slower.
Public source
Publisher name
Public post
If you’re using a general-purpose LLM for translation, quality equals latency. If you’re using a specialized model like DeepL, it doesn’t. OpenAI GPT-5.2 is 4x slower th…
Company
DeepL
Building the AI language layer for global business
- Industry
- Software Development
- Location
- Cologne, DE
- Company size
- 1,001–5,000 employees
About DeepL
DeepL is the AI Research company building an AI language layer for global business. Our work tackles the most difficult challenges in multilingual communication. This includes innovating new approaches to translation quality and determinism, and enabling real-time, live speech-to-speech translation that matches an original speaker’s tone, accent and style. We develop this research into end-to-end products that are designed to meet the needs of global business. DeepL’s Translator platform delivers the top-performing translations in blind tests with human linguists. Our real-time Voice translations are rated as market-leading in independent tests. Our DeepL API delivers translation everywhere, with an unmatched combination of quality and reduced latency.
See moreLatest activity
Latest activity from DeepL
20 signals
People
DeepL is hiring for a Senior or Mid-level Site Reliability Engineer in the UK based in London, requiring knowledge of Helm from Kustomize, Thanos from Prom, and SLIs from SLOs.
Research & Knowledge
DeepL published a DeepL Journal du Net article discussing the European approach to integrating AI into business processes after an initial phase of experimentation.
Partnerships
DeepL partnered with MACNICA to provide DeepL Voice for Conversations free to local governments, international associations, and nonprofits across the prefecture
Discover more
Similar signals
Similar public activity from other companies.
Research & Knowledge
Google Cloud reported that Artificial Analysis clocked Gemini 3.8 Flash High at 333.6 tokens per second and Opus 5 Max at 55 tokens per second
Research & Knowledge
Vals AI
Vals AI reports that Gemini 3.8 Flash scores 62.3% on the Vals Index, fifth behind Claude Fable 5.1, Claude Opus 5, Claude Fable 5, and GPT-5.6 Sol.
Products & Services
Google DeepMind
Google DeepMind released Claude Opus 5, which beats Gemini 3.8 Flash on 8 of 14 benchmark evaluations at 85% lower cost.
Research & Knowledge
Perplexity
Perplexity evaluated GPT-6 Astra on WANDR, scoring 27.0% higher than Opus 5 at 3.3% higher cost.
Research & Knowledge
AlphaSense