Inflection-2.5
Near GPT-4 performance on 60% less of compute for training. Inflection-2.5 approaches GPT-4’s performance, but used only 40% of the amount of compute for training.
Reactions (1)
Similar products
by description similarity
GPT‑5.4
openai.comOpenAI's most efficient model: less tokens, more clarity
New LLM outperforming GPT-3.5
news.ycombinator.comRefuel LLM (84.2%) outperforms trained human annotators (80.4%), GPT-3-5-turbo (81.3%), PaLM-2 (82.3%) and Claude (79.3%) across a benchmark of 15 text labeling datasets.
GPT-5.1 Instant and Thinking
openai.comSmarter, more conversational ChatGPT with new personalities
GPT-4 level summarization at 63% less cost and 11 times the speed
blog.langchain.devLearn how fine-tuning ChatGPT with synthetic data surpasses GPT-4 summarization while cutting costs by 63% and boosting speed 11x using LangSmith.
GPT-4
openai.comLLM that exhibits human-level performance