441e85161ClaimNous Research trained Arete-27B using supervised fine-tuning on approximately 53 million tokens over two epochs without reinforcement learning or preference optimizationAgreeDisagree
441e85161ClaimThe training dataset for Arete-27B contains 34,878 prompts across seven value categories, with 29,236 accepted teacher answers forming the fine-tuning splitAgreeDisagree
441e85161ClaimInkling 2 Small features 594 billion total parameters with a sparse architecture that activates 17 billion parameters per tokenAgreeDisagree
441e85161ClaimPipecat is an open-source framework that orchestrates voice and video streams while linking speech-to-text, large language models, and text-to-speech components to minimize latency in AI applications.AgreeDisagree
441e85161ClaimBrave Software cited in a blog post that Anthropic used its Search API to train the Claude 2 artificial intelligence model.AgreeDisagree
441e85161ClaimDeepgram expanded its Nova-3 speech recognition model to support 55 languages as part of its product strategy.AgreeDisagree
441e85161ClaimBaidu pre-trained ERNIE 5.5 Flash Base on 33.5 trillion tokens using the Muon optimizer with a 128,000-token training context windowAgreeDisagree
441e85161ClaimBaidu released ERNIE 5.5 Flash Base, a 280-billion-parameter text-only sparse Mixture-of-Experts foundation model that activates 14 billion parameters per tokenAgreeDisagree
441e85161ClaimERNIE 5.5 Flash Base utilizes a Multi-Router Latent MoE architecture containing 512 routed experts running in a 2048-dimensional latent space alongside one shared expertAgreeDisagree
441e85161ClaimERNIE 5.5 Flash Base implements ECHO speculative decoding, which reuses the final backbone layer as a draft model adding 0.36 billion parameters excluded from headline countsAgreeDisagree
441e85161ClaimERNIE 5.5 Flash Base features a hybrid attention mechanism across 43 layers, using Heavily Compressed Attention in 37 layers and Multi-head Latent Attention with sparse attention in 6 layersAgreeDisagree
441e85161ClaimERNIE 5.5 Flash Base incorporates a tokenizer with 201,216 entries that groups numbers right-to-left in chunks of up to three digitsAgreeDisagree
441e85161ClaimRicursive Intelligence pretrained Ripple-65B Preview from scratch on 12 trillion text and code tokens at an 8,000 sequence lengthAgreeDisagree
441e85161ClaimRicursive Intelligence reported requiring roughly 70 times fewer reinforcement learning rollouts for Ripple-65B Preview than the 100 million rollouts reported for Reflection BeamAgreeDisagree
441e85161ClaimRicursive Intelligence reached target reinforcement learning performance for Ripple-65B Preview using 1.4 million rolloutsAgreeDisagree
441e85161ClaimRicursive Intelligence's Ripple-65B Preview features a dense Transformer architecture with 65 billion parameters across 80 layers and a model dimension of 5120AgreeDisagree
441e85161ClaimPretraining of Ripple-65B Preview achieved roughly 40% model FLOPs utilization on NVIDIA hardware without custom kernels or quantization during trainingAgreeDisagree
441e85161ClaimThe experimental NVFP4 checkpoint of Ricursive Intelligence's Ripple-65B Preview quantizes MLP layers while keeping attention in BF16 to target local hardware such as NVIDIA DGX SparkAgreeDisagree
441e85161ClaimStartup Ricursive Intelligence was founded by CEO Anna Goldie and CTO Azalia Mirhoseini, who previously helped create Google's AlphaChip chip design softwareAgreeDisagree
441e85161ClaimRicursive Intelligence shifted from Olmo 3's Llama-like layout to a Gemma 3N-like design for Ripple-65B Preview to improve model stabilityAgreeDisagree