model releases (L1): 144 curated resources
model releases (L1)
144 resources · page 1 of 5
π0.7 (Physical Intelligence)
Physical Intelligence blog post on π0.7, a steerable robot foundation model with cross-robot transfer.
AGIBOT BFM / GCFM
by AgiBot
AGIBOT announcement (April 2026) of new embodied AI robots and foundation models.
AliceAI-Foundation-80B-A3B-Base (Yandex)
by Yandex
Hugging Face model card for yandex/AliceAI-Foundation-80B-A3B-Base, Yandex's 80B-total, 3B-active base model, with weights, usage and license details.
Apertus-v1.5-70B
Hugging Face model card for swiss-ai/Apertus-v1.5-70B, Swiss AI's 70B fully open multilingual model, with weights, usage and license details.
Baichuan-M2-32B (Baichuan AI)
by Baichuan Intelligent Technology
Baichuan-M2-32B GitHub repo: open 32B medical language model with training and evaluation details.
Baichuan-M3 (Baichuan AI)
by Baichuan Intelligent Technology
Baichuan-M3-235B GitHub repo: open 235B medical model trained for clinical inquiry and decision support.
Beam 1.0 (Reflection AI)
by Reflection
Hugging Face model card for reflection/Beam-1.0-501B-A23B, Reflection AI's 501B-total, 23B-active open mixture-of-experts model, with weights, usage and license details.
BGE-M3 (BAAI)
by Beijing Academy of Artificial Intelligence
Hugging Face model card for BAAI BGE-M3: multilingual embedding model supporting dense, sparse and multi-vector retrieval.
Claude Fable 5
by Anthropic
Anthropic announcement of Claude Fable 5 and Claude Mythos 5 with capabilities and safety information.
Claude Fable 5.1
by Anthropic
Claude Platform page for Claude Fable 5.1: model ID, 1M-token context, pricing and changes from Fable 5.
Claude Fable 5 (high)
by Anthropic
Claude Platform page for Claude Fable 5: model ID, context window, pricing and effort settings such as high.
Claude Opus 4.6
by Anthropic
Anthropic announcement of Claude Opus 4.6 with coding, computer-use and tool-use results.
Claude Opus 4.7
by Anthropic
Anthropic announcement of Claude Opus 4.7 with software engineering improvements and evaluations.
Claude Opus 4.8
by Anthropic
Anthropic announcement of Claude Opus 4.8 with coding, agentic and long-running task results.
Claude Opus 5
by Anthropic
Anthropic announcement of Claude Opus 5 with long-running agent, coding and professional-work results.
Claude Sonnet 5
by Anthropic
Anthropic announcement of Claude Sonnet 5 with agentic coding capabilities and benchmark results.
Codestral 25.08
by Mistral AI
Mistral documentation page for Codestral 25.08, its code generation model, with specs and API usage.
Oasis
by oasis-model
Oasis project page (Decart and Etched): real-time interactive world model that generates a playable game from user input.
DeepSeek-OCR-2
by DeepSeek
DeepSeek-OCR-2 GitHub repo: document OCR vision-language model with code and weights.
DeepSeek-V4.1-Flash
by DeepSeek
DeepSeek announcement of DeepSeek-V4.1-Flash, a small model in its new architecture with visual understanding.
DeepSeek V4 Pro 0813
Artificial Analysis page for DeepSeek V4 Pro 0813 (DeepSeek): intelligence index, price, output speed and latency compared with other models.
EmbeddingGemma-300m
by Google
Hugging Face model card for google/embeddinggemma-300m, Google's 300M-parameter text embedding model, with weights, usage and license details.
ERNIE 5.1 (Baidu)
by Baidu
Baidu ERNIE blog: releases and technical notes for ERNIE models, including ERNIE 5.1.
EuroLLM-22B-Instruct-2512 (UTTER, EU-funded)
Hugging Face model card for utter-project/EuroLLM-22B-Instruct-2512, the EU-funded UTTER project's 22B multilingual European instruction model, with weights, usage and license details.
EXAONE 4.5 33B (LG)
Hugging Face model card for LG EXAONE 4.5 33B: weights, usage and evaluation results.
Falcon-H1R-7B (TII)
by Technology Innovation Institute
Hugging Face model card for tiiuae/Falcon-H1R-7B-FP8, TII's 7B Falcon-H1R hybrid reasoning model in FP8, with weights, usage and license details.
Falcon Perception (TII)
by Aviraj Bevli et al. (Technology Innovation Institute)
arXiv paper (PDF, March 2026) describing TII's Falcon Perception model.
Figure Helix
by Figure AI
Figure AI post introducing Helix, a vision-language-action model for humanoid robot control.
G9v3-39A5B
Artificial Analysis page for G9v3-39A5B (AI9Stars): intelligence index, price, output speed and latency compared with other models.
G9v3-3B
Artificial Analysis page for G9v3-3B (AI9Stars): intelligence index, price, output speed and latency compared with other models.
Gemini 3.1 Pro preview
by Google
Gemini API model page for Gemini 3.1 Pro preview: model code, capabilities, token limits and versions.
Gemini 3.5 Flash
by Google
Google blog announcement of the Gemini 3.5 model series, including Gemini 3.5 Flash.