Cam SmithAI Studio

Computational Models

Explore model releases

A model is an identifiable computational design, not the whole application around it. Start with its designed function, then inspect variants, artifacts, hosted offerings, and evidence.

Read the boundaries

A model’s contribution is specific.

An encoder can supply representations used by retrieval, while the surrounding index and query logic perform retrieval. A generative model can propose text, while a separate runtime handles tools and permissions.

67 cited releases

Model releases

Output modality describes what a model returns. Use Specific function to distinguish video generation from segmentation masks, or speech from music and sound effects.

Designed contribution
Specific function
Input modality
Output modality
Access channel
Weight access
Deployment
Availability
More filters
Specialization
Technical form
Developer
Lifecycle
Family
Clear filters

Showing all 67 model releases.

  • Model release · Beijing Academy of Artificial Intelligence

    BGE English v1.5

    Produces numerical representations of English text for similarity-based applications.

    Designed contribution
    representation
    Input
    text
    Output
    numeric-vector
    Output form
    embedding vector
    Access paths

    Last verified:

  • Model release · Beijing Academy of Artificial Intelligence

    BGE Reranker v2 M3

    Scores how relevant a candidate passage is to a query, supporting a second stage of text retrieval.

    Designed contribution
    retrieval ranking
    Input
    text
    Output
    numeric-scalar
    Output form
    relevance score
    Access paths

    Last verified:

  • Model release · Anthropic

    Claude Fable 5.1

    Demanding reasoning and long-horizon agentic work with text responses.

    Designed contribution
    inference planning
    Input
    image, text
    Output
    text
    Output form
    generated text response
    Access paths

    Last verified:

  • Model release · Anthropic

    Claude Haiku 4.5

    Fast text and image understanding with text responses for routine tasks.

    Designed contribution
    inference planning
    Input
    image, text
    Output
    text
    Output form
    generated text response
    Access paths

    Last verified:

  • Model release · Anthropic

    Claude Opus 5.5

    Long-running agentic coding and knowledge work with text responses.

    Designed contribution
    inference planning
    Input
    image, text
    Output
    text
    Output form
    generated text response
    Access paths

    Last verified:

  • Model release · Anthropic

    Claude Sonnet 5.5

    General reasoning, coding, and knowledge tasks with an emphasis on speed and capability.

    Designed contribution
    inference planning
    Input
    image, text
    Output
    text
    Output form
    generated text response
    Access paths

    Last verified:

  • Model release · Cohere

    Cohere Embed v4.0

    Produces embeddings from text, images, or mixed text and image inputs through a documented API.

    Designed contribution
    representation
    Input
    image, text
    Output
    numeric-vector
    Output form
    embedding vector
    Access paths

    Last verified:

  • Model release · Cohere

    Command A (03-2025)

    Generates text through Cohere’s hosted chat service using the exact command-a-03-2025 model ID.

    Designed contribution
    generation transformation
    Input
    text
    Output
    text
    Output form
    generated text
    Access paths

    Last verified:

  • Model release · ElevenLabs

    Eleven Music v2.5

    Generate music from text prompts or composition plans; the web interface also accepts an audio style reference.

    Designed contribution
    generation transformation
    Input
    text
    Output
    audio
    Output form
    generated music track
    Access paths

    Last verified:

  • Model release · ElevenLabs

    Eleven v4

    Generate expressive multilingual speech from text.

    Designed contribution
    generation transformation
    Input
    text
    Output
    audio
    Output form
    synthesized speech
    Access paths

    Last verified:

  • Model release · Black Forest Labs

    FLUX.2 [klein] 9B

    Generate images from text and edit images with multiple references using FLUX.2 [klein] 9B.

    Designed contribution
    generation transformation
    Input
    image, text
    Output
    image
    Output form
    generated image, multi-reference edited image
    Access paths

    Last verified:

  • Model release · Google

    Gemini 3 Pro Image

    Generate and edit professional visual assets from prompts and reference images.

    Designed contribution
    generation transformation
    Input
    image, text
    Output
    image, text
    Output form
    generated or edited image, supporting text
    Access paths

    Last verified:

  • Model release · Google

    Gemini 3.1 Flash Image

    Generate and conversationally edit images from text, image, or video context.

    Designed contribution
    generation transformation
    Input
    image, text, video
    Output
    image, text
    Output form
    generated or edited image, supporting text
    Access paths

    Last verified:

  • Model release · Google

    Gemini 3.8 Flash

    Fast multimodal text response generation for coding and agentic tasks.

    Designed contribution
    inference planning
    Input
    audio, image, text, video
    Output
    text
    Output form
    generated text response
    Access paths

    Last verified:

  • Model release · Google

    Gemini Omni Flash 1.1

    Generate, edit, extend, and interpolate short videos from text, image, or video input.

    Designed contribution
    generation transformation
    Input
    image, text, video
    Output
    video
    Output form
    generated or edited video clip
    Access paths

    Last verified:

  • Model release · Google

    Gemma 2 instruction-tuned

    Google instruction-tuned Gemma 2 text model line, with distinct size variants and gated artifacts.

    Designed contribution
    generation transformation
    Input
    text
    Output
    text
    Output form
    generated text
    Access paths

    Last verified:

  • Model release · OpenAI

    GPT-6 Astra

    Reasons over text and images and generates text through OpenAI's hosted API under the exact gpt-6-astra model ID.

    Designed contribution
    inference planning
    Input
    image, text
    Output
    text
    Output form
    generated text, reasoned text response
    Access paths

    Last verified:

  • Model release · OpenAI

    GPT-6 Luna

    Focused, high-volume reasoning and text generation through OpenAI API.

    Designed contribution
    inference planning
    Input
    image, text
    Output
    text
    Output form
    generated text response
    Access paths

    Last verified:

  • Model release · OpenAI

    GPT-6.1 Sol

    General reasoning, coding, and document work through a current OpenAI API model.

    Designed contribution
    inference planning
    Input
    image, text
    Output
    text
    Output form
    generated text response
    Access paths

    Last verified:

  • Model release · OpenAI

    GPT-Audio-1.5

    Accept text or audio and produce spoken or text responses through Chat Completions.

    Designed contribution
    generation transformation
    Input
    audio, text
    Output
    audio, text
    Output form
    generated spoken response, text response
    Access paths

    Last verified:

  • Model release · OpenAI

    GPT-Image-2.5 Flare

    Generate or edit images from text and image input for fast everyday visual work.

    Designed contribution
    generation transformation
    Input
    image, text
    Output
    image
    Output form
    generated or edited image
    Access paths

    Last verified:

  • Model release · OpenAI

    GPT-Image-2.5 Sunburst

    Generate images from text prompts and edit supplied images with an emphasis on precise edits.

    Designed contribution
    generation transformation
    Input
    image, text
    Output
    image
    Output form
    generated or edited image
    Access paths

    Last verified:

  • Model release · OpenAI

    GPT-Live 1

    Hold full-duplex spoken conversations with simultaneous listening and speech output.

    Designed contribution
    generation transformation
    Input
    audio, text
    Output
    audio, text
    Output form
    streaming spoken response, text response
    Access paths

    Last verified:

  • Model release · IBM Research

    IBM pos-egnn v1-6M

    Predicts energy, per-atom forces and stress for atomic structures using a position-based graph neural network.

    Designed contribution
    prediction estimation
    Input
    graph, spatial
    Output
    numeric-scalar, spatial
    Output form
    per-atom force vectors, potential energy, stress tensor when enabled
    Access paths

    Last verified:

  • Model release · Kuaishou

    Kling Video 3.0

    Generate and edit multi-shot video from prompts and visual references, with native audio.

    Designed contribution
    generation transformation
    Input
    image, text, video
    Output
    audio, video
    Output form
    generated video with native audio
    Access paths

    Last verified:

  • Model release · hexgrad

    Kokoro v1.0

    Turns supplied text into synthesized speech using a downloadable v1.0 checkpoint.

    Designed contribution
    generation transformation
    Input
    text
    Output
    audio
    Output form
    speech waveform
    Access paths

    Last verified:

  • Model release · Google

    Lyria 3.5

    Generate songs and instrumentals from text or image direction.

    Designed contribution
    generation transformation
    Input
    image, text
    Output
    audio
    Output form
    generated full-length song
    Access paths

    Last verified:

  • Model release · Midjourney

    Midjourney V8.2

    Generate images from text and image prompts using Midjourney’s current V8.2 model.

    Designed contribution
    generation transformation
    Input
    image, text
    Output
    image
    Output form
    generated image
    Access paths

    Last verified:

  • Model release · Mistral AI

    Mistral Small 4

    Instruction following, reasoning, and coding in one hosted Mistral model.

    Designed contribution
    inference planning
    Input
    text
    Output
    text
    Output form
    generated text response
    Access paths

    Last verified:

  • Model release · Google Research / NeuralGCM team

    NeuralGCM v1 1.4° Stochastic

    Simulates atmospheric evolution using a released hybrid machine-learning and physics checkpoint.

    Designed contribution
    simulation
    Input
    spatial, time-series
    Output
    spatial, time-series
    Output form
    simulated atmospheric state trajectory
    Access paths

    Last verified:

  • Model release · Namasivayam Kalithasan et al. (IIT Delhi)

    Neuro-symbolic Robot Manipulation (ICRA 2023)

    Research hybrid that grounds a natural-language instruction in an initial depth-image scene to produce a symbolic manipulation program and pose sub-goals.

    Designed contribution
    inference planning
    Input
    image, text
    Output
    spatial, symbolic-structured
    Output form
    grounded manipulation program, object and target-pose sub-goals
    Access paths

    Last verified:

  • Model release · Mobile Sensing and Ubiquitous Computing Laboratory

    NormWear

    Encodes multichannel physiological wearable sensor signals into reusable signal representations.

    Designed contribution
    representation
    Input
    sensor, time-series
    Output
    numeric-vector
    Output form
    per-channel signal embeddings, per-patch signal embeddings
    Access paths

    Last verified:

  • Model release · OpenVLA collaboration

    OpenVLA 7B

    Predicts robot manipulation actions from a camera image and an English-language instruction.

    Designed contribution
    control action
    Input
    image, text
    Output
    numeric-vector
    Output form
    normalized 7-DoF robot action
    Access paths

    Last verified:

  • Model release · NVIDIA

    Parakeet TDT 0.6B v2

    Transcribes English speech audio into text with punctuation, capitalization, and timestamp support.

    Designed contribution
    perception
    Input
    audio
    Output
    text
    Output form
    speech transcript, word timestamps
    Access paths

    Last verified:

  • Model release · Qwen

    Qwen3.5-397B-A17B

    Vision-language reasoning and text generation with Qwen’s 397B-A17B release.

    Designed contribution
    inference planning
    Input
    image, text
    Output
    text
    Output form
    generated text response
    Access paths

    Last verified:

  • Model release · Qwen

    Qwen3.5-9B

    Vision-language reasoning and text generation with Qwen’s 9B release.

    Designed contribution
    inference planning
    Input
    image, text
    Output
    text
    Output form
    generated text response
    Access paths

    Last verified:

  • Model release · Daniel Gnad and Álvaro Torralba

    Recharging Robots (IPC 2023 PDDL domain)

    A pinned engineered PDDL domain describing robots, battery use, guarding, and action transitions for IPC 2023 planning tasks.

    Designed contribution
    representation
    Input
    symbolic-structured
    Output
    symbolic-structured
    Output form
    PDDL action and transition definitions
    Access paths

    Last verified:

  • Model release · Runway

    Runway Gen-4.5

    Generate video clips from text prompts or image-plus-text direction.

    Designed contribution
    generation transformation
    Input
    image, text
    Output
    video
    Output form
    generated video clip
    Access paths

    Last verified:

  • Model release · Meta

    SAM 2.1 Hiera Large

    Segments prompted objects in images and follows their masks through video frames.

    Designed contribution
    perception
    Input
    image, spatial, video
    Output
    image, video
    Output form
    segmentation masks, tracked video masks
    Access paths

    Last verified:

  • Model release · ByteDance Seed

    Seedance 2.5

    Generate and extend up to 30-second audio-video clips from multimodal creative direction.

    Designed contribution
    generation transformation
    Input
    audio, image, text, video
    Output
    audio, video
    Output form
    generated audio-video clip
    Access paths

    Last verified:

  • Model release · Google

    ShieldGemma 1

    Google text safety moderation models in 2B, 9B, and 27B sizes, built on Gemma 2.

    Designed contribution
    prediction estimation
    Input
    text
    Output
    text
    Output form
    Yes/No text classification signal
    Access paths

    Last verified:

  • Model release · Hugging Face

    SmolVLM Instruct

    Answers visual questions and describes images in text with a compact vision-language model.

    Designed contribution
    perception
    Input
    image, text
    Output
    text
    Output form
    image description, visual question answer
    Access paths

    Last verified:

  • Model release · Suno

    Suno v6

    Generate and edit songs from prompts or multimodal references using Suno’s v6 model.

    Designed contribution
    generation transformation
    Input
    audio, image, text, video
    Output
    audio
    Output form
    generated song
    Access paths

    Last verified:

  • Model release · Prior Labs

    TabPFN v2 Classifier

    Predicts class labels from small tabular datasets using a pretrained tabular model.

    Designed contribution
    prediction estimation
    Input
    tabular
    Output
    tabular
    Output form
    class predictions
    Access paths

    Last verified:

  • Model release · OpenAI

    text-davinci-003

    Historical OpenAI instruction-following text-completion model whose public API access ended in 2024.

    Designed contribution
    generation transformation
    Input
    text
    Output
    text
    Output form
    text completion
    Access paths

    Last verified:

  • Model release · Google Research

    TimesFM 2.0 500M PyTorch

    Forecasts future values of a univariate time series from its historical observations.

    Designed contribution
    prediction estimation
    Input
    time-series
    Output
    time-series
    Output form
    point forecast sequence
    Access paths

    Last verified:

  • Model release · Google

    Veo 3.1 Preview

    Generate short videos with natively generated audio from text or image direction.

    Designed contribution
    generation transformation
    Input
    image, text
    Output
    audio, video
    Output form
    generated video with native audio
    Access paths

    Last verified:

  • Model release · Mistral AI

    Voxtral TTS

    Synthesize speech from text, including voice-conditioned delivery.

    Designed contribution
    generation transformation
    Input
    audio, text
    Output
    audio
    Output form
    synthesized speech
    Access paths

    Last verified:

  • Model release · Wan Team

    Wan2.1 T2V 1.3B

    Generates short video from text prompts using a downloadable 1.3B checkpoint.

    Designed contribution
    generation transformation
    Input
    text
    Output
    video
    Output form
    generated video
    Access paths

    Last verified:

  • Model release · Wan-AI

    Wan2.2 T2V-A14B

    Generate 5-second video clips from text prompts using publisher-distributed Wan2.2 T2V weights.

    Designed contribution
    generation transformation
    Input
    text
    Output
    video
    Output form
    generated video clip
    Access paths

    Last verified: