Sublime
An inspiration engine for ideas
DeFi AI
Gwindorian • 6 cards
Closing the Gap to Closed Source LLMs – Open-Sourcing 70B Abacus Giraffe!
The best-performing open-source model on MT-bench in key categories
We are super excited to be open-sourcing our best model for Enterprise AI use cases - the 70B 32K context length Abacus Giraffe model!... See more
would love if you could bench deepseek r1 distills. you can do this with llama.cpp (more accurate) or https://t.co/POZPRrv92O. also would be interesting if there is any difference in speed between gguf and mlx runtimes.
if your time allows, i'd love to see numbers for 14b, 32b, 70b models, at different quantizations (say, 4, 8 bits)... See more
bantegx.com
GPT-5 Pro Is Now On LiveBench
GPT-5 Pro and High are both top of LiveBench and are similar to each other
Overall, GPT-5 thinking models are excellent at problem-solving, but can become unbearably slow as they occasionally enter a death loop.
And yeah, we haven't... See more
We’re releasing PaperBench, a benchmark evaluating the ability of AI agents to replicate state-of-the-art AI research, as part of our Preparedness Framework.
Agents must replicate top ICML 2024 papers, including understanding the paper, writing code, and executing experiments. https://t.co/CvYcDdk0nI
🚀Introducing The LLM Inference Provider Leaderboard https://t.co/f1LzVce9yb - a live-updated, unbiased eval of API Inference products.
Featuring: @abacusai, @anyscalecompute, @DeepInfra, @DecartAI, @FireworksAI_HQ, @LeptonAI, @togethercompute, @perplexity_ai, @replicate, as well as @OpenAI and @AnthropicAI... See more
Martianx.com
