DeepSeek V3.2 Exp is an experimental iteration built on the V3.1 “Terminus” lineage and intended as a stepping stone toward DeepSeek’s next-generation architecture. It focuses on efficiency (lower training/inference cost) for long-context workloads while aiming to preserve V3-class output quality. The company positioned it as an intermediate release meant to validate new attention and routing […]
Model Type: Chat
Anthropic’s Claude Sonnet 4.5 is a 2025 update focused on long-duration autonomous work, stronger tool/computer use, tighter safety/alignment, and richer developer features (context editing, memory, in-chat code execution and file creation). Benchmarks and vendor tests emphasize large gains on coding and “agentic” tasks (e.g., Anthropic reports ~30 hours of continuous autonomous coding in internal tests and big jumps on OS/tool-use benchmarks).
Model Type: LLm
Grok-4-Fast is xAI’s new cost-efficient reasoning model designed to make high-quality reasoning and web search capabilities cheaper and faster for both consumer and developer use. xAI positions it as a frontier offering that preserves Grok-4’s benchmark performance while improving token efficiency, and ships two variants tuned for either reasoning or non-reasoning workloads.
Model Type: Chat
grok-code-fast-1 is xAI’s speed-focused, cost-efficient agentic coding model designed to power IDE integrations and automated coding agents. It emphasizes low latency, agentic behaviors (tool calls, stepwise reasoning traces), and a compact cost profile for day-to-day developer workflows.
Model Type: Code
Runway/upscale_v1 is Runway’s production-targeted video upscaler model (model id upscale_v1) designed to increase video resolution by 4× up to 4K (capped at 4096 pixels on a side). It’s available through Runway’s API surface and is also packaged on third-party model hosting marketplaces (e.g., CometAPI).
Model Type: Video
Gen-4 Image is Runway’s flagship multimodal image-generation model in the Gen-4 family that supports prompted generation plus visual references (you can “@mention” reference images) to produce highly controllable, stylistically consistent outputs for image and image→video pipelines.
Model Type: Image Generation
Runway Gen-4 Aleph(model id /gen4_aleph) is Runway’s in-context, video-to-video model that extends the Gen-4 family with powerful video editing, shot continuation and view-synthesis capabilities. In plain terms: Aleph can take an input clip and perform complex edits — add/remove/replace objects, relight, restyle, generate novel camera angles, and even generate the “next shot” in a sequence — driven by text prompts and optional reference images. This release is presented as a major step toward coherent, multi-shot video generation and in-context editing.
Model Type: Video
Act-Two is Runway’s next-generation AI performance capture and character animation tool: it ingests a short driving performance (a webcam or phone video of someone acting a scene) plus a character reference (image or video) and generates an animated character performance that transfers body, facial expression and hand motion to the character. Act-Two is offered inside Runway’s web product and as a model available through API ecosystem.
Model Type: Video
Seedream 4.0 (model identifier doubao-seedream-4-0-250828) is ByteDance’s newest image-creation foundation model that unifies text→image generation and image editing in a single architecture. The model was announced in early September 2025 and is positioned for high-resolution (up to 4K) generation, multi-reference workflows, and fast inference for production use.
Model Type: Image Generation
Hailuo 02 API is an asynchronous, HTTP‑based service from MiniMax that lets developers generate professional‑grade, cinematic videos from either text or image prompts.
Model Type: Video
