shipfeedAI news, curated daily

23:06:32 CET
20 MAY23:06:32shipfeed
pull to refreshlast sync
Just in — 30 new
§ feed · storyline

Inflection-2.5 at 94% of GPT4, and Pi at 6m MAU

Inflection launches model 2.5, reaching 94% of GPT-4 average performance using 40% of the training compute, while Pi reports 6 million monthly active users growing at 10% weekly.

Mar 8 · · primary fetch1 sourceupdated Mar 8 ·

Mustafa Suleyman announced Inflection 2.5, which achieves more than 94% the average performance of GPT-4 despite using only 40% the training FLOPs. Pi's user base is growing about 10% weekly, with new features like realtime web search. The community noted similarities between Inflection 2.5 and Claude 3 Sonnet. Claude 3 Opus outperformed GPT-4 in a 1.5:1 vote and is now the default for Perplexity Pro users. Anthropic added experimental tool calling support for Claude 3 via LangChain.

LlamaIndex released LlamaParse JSON Mode for structured PDF parsing and added video retrieval via VideoDB, enabling retrieval-augmented generation (RAG) pipelines. A paper proposed knowledge-augmented planning for LLM agents. New benchmarks like TinyBenchmarks and the Yi-9B model release show strong code and math performance, surpassing Mistral.

read full article on news.smol.ai
§ sources1 publication · timeline below
  1. news.smol.aiInflection-2.5 at 94% of GPT4, and Pi at 6m MAUprimary
Inflection-2.5 at 94% of GPT4, and Pi at 6m MAU · shipfeed