#DeepSeek
2 articles tagged DeepSeek
Tagged articles
Articles tagged DeepSeek, newest first.
AI PerformanceDeepSeek V4.1 Flash Hits 20.7% on Plumbing Takeoff, Far Behind Claude Opus 5.5
DeepSeek V4.1 Flash hit 20.7% on a plumbing takeoff, Claude Opus 5.5 hit 42.7%. On 169 drawing questions it scored 62.7% overall and just 9.1% on measurements. API price of $0.30 per million input tokens, about one tenth of rival list prices. A cost report test turned a $250,000 forecast loss into a $1.03 million profit. Verify every takeoff count and schedule link before using low-cost model output.
AI NewsDeepSeek Opens DeepInfer Serving Engine Alongside MIT-Licensed V4.1-Flash
DeepSeek open-sourced DeepInfer, the engine behind its API, days after V4.1-Flash. Heavy coding day: about $2.10 off-peak on V4.1-Flash versus $30 on Gemini 4 Argon. V4.1-Flash scores 42.3% on Terminal-Bench 4.0 versus 66.4% for Claude Opus 5.5. DeepInfer targets Blackwell-class GPUs and Ascend 950 NPUs, offline inference only. Use Flash for routine coding in Claude Code, frontier models for hard refactors.
Get new articles by RSS or podcast
Add the RSS URL to your RSS reader or the podcast URL to your podcast app to get every new article as soon as it is published. The podcast reads out the headline and key points.
How to subscribe