Runway

  • 数据已交叉验证，Runway 2026 定价改档关键信息（Unlimited → Max、Gen-4.5 上线）已确认。下面这张按你锁定的表格骨架铺出：

    Runway: The Filmmaker’s AI Video Studio — Gen-4.5 / Gen-4 / Turbo Multi-Model Router, World-Consistent Characters, Aleph In-Video Editing, Act-Two Performance Capture, and Credit-Metered 4K Generation

    • Name: Runway (RunwayML / Runway AI)
    • Official Website: https://runwayml.com
    • Short Description: Browser-based AI video creation suite spanning text/image/video-to-video generation (Gen-4.5, Gen-4, Gen-4 Turbo), image generation, consistent character/location/object world-building, Aleph prompt-based in-video editing, Act-Two/Act-One performance capture, Motion Brush, Director Mode camera control, 4K upscaling, video editor + green screen + frame interpolation, plus a multi-model router (Veo 3.1, Kling 3.0 Pro, Seedance 2.0, Nano Banana Pro, FLUX.2, Seedream 5.0). Credit-metered freemium; founded 2018 (NYC, Cristóbal Valenzuela, Anastasis Germanidis, Alejandro Matamala). Not a text-to-app builder, not an avatar bench (HeyGen/Synthesia), not a repurpose tool (Lumen5/Pictory).

    1. One-Line Positioning

    Runway = the filmmaker/agency-grade AI video workbench where image-to-video is the native idiom​ (drop a still → animate with cinematic motion + camera control), built around world consistency​ (one reference image locks a character/object/location across every shot with no fine-tuning), extended by Aleph (edit-after-generate), Act-Two (webcam performance → character), and a model router that lets one subscription reach Gen-4.5, Veo 3.1, Kling, Seedance. Fits ai-productivity-data; pipeline_stage=none (prompt → generate → extend → edit → upscale → comp). 2026 trap: credits reset monthly with no roll-over on Free/Standard/Pro; failed generations still cost credits; Gen-4.5 ≈ 25 credits/sec means a Pro month (2,250) buys only ~90 sec of flagship footage.

    2. Core Features (Table)

    Layer Capability Notes
    Generation​ Text-to-Video, Image-to-Video (native), Video-to-Video (style transfer / restyle) 5 s / 10 s clips; extend via stitching
    Models​ Gen-4.5 (text-to-video flagship, ~25 cr/s), Gen-4 (image-to-video, ~12 cr/s), Gen-4 Turbo (~5 cr/s, ~60% cheaper, <15 s render, 60 fps) Multi-model router also exposes Veo 3.1, Kling 3.0 Pro, Seedance 2.0, Nano Banana Pro, FLUX.2, Seedream 5.0
    Image​ Text-to-Image (Gen-4 image, ~5–10 cr), references, 4K upscaling Included on all paid tiers
    Consistency​ Consistent Characters / Objects / Locations across scenes from 1 reference image; multi-angle coverage No fine-tuning or training required
    Editing​ Aleph​ — prompt-based in-video edit (remove object, extend scene, change bg, add element) with AI masking Post-generation editing paradigm
    Performance​ Act-Two​ (was Act-One) — webcam facial + body capture → transfer to AI character; custom voices for lip-sync/TTS ~80–120 credits per use
    Motion / Camera​ Motion Brush 3.0 (region-painted independent motion vectors), Director Mode (node-based camera: zoom/pan/tilt/truck/orbit + keyframes) Best-in-class creative control
    Multi-shot​ Multi-Shot Video / narrative sequences, temporal attention layers (no flicker) Up to 16 s continuous; extends via stitching
    VFX​ GVFX (generative VFX sitting beside live action), green screen, frame interpolation, inpainting Studio-grade compositing
    Editor​ Browser video editor, 3 projects (Free) → unlimited, assets + timeline, 4K export (Pro+) Not a Premiere/DaVinci replacement
    Audio​ Generative Audio, Text-to-Speech (custom voices on Pro+), Audio Apps Audio is additive, not core
    API​ Python / Node SDK, REST — programmatic Gen-4 + tools $0.01/credit retail; Enterprise for team APIs
    Collaboration​ Workspaces, teamspaces, workspace analytics (Enterprise), brand/style assets Up to 5 (Standard) / 10 (Pro/Max) seats per workspace

    3. Pricing (2026-08, USD — Credit-Metered Freemium)

    Verified against runwayml.com/pricing; 2026 plan rename: old Basic→Standard, Standard→Pro, Pro→Max, Unlimited discontinued​ (existing Unlimited users migrate Sep 1 2026; Max = replacement).

    Tier $/mo (monthly) $/mo (annual, −20%) Credits/mo What’s In
    Free​ $0 $0 125 one-time​ (no renewal) Gen-4 Turbo only, watermark, 720p, 3 editor projects, 5 GB, commercial use allowed
    Standard​ $15 $12 625/mo All models (Gen-4.5, Veo 3.1, Kling 3.0 Pro, FLUX.2, Nano Banana Pro…), 4K upscaling, no watermark, 100 GB
    Pro​ $35 $28 2,250/mo Everything in Standard + custom voices for lip-sync/TTS, 500 GB, best models (Gen-4.5, Seedance 2.0, Kling 3.0…), priority generation
    Max​ $95 $76 9,500/mo (raised from 6,750 on old Pro) Everything in Pro + credits roll over 1 month​ + first access to newest models, highest generation volume
    Enterprise​ Custom Custom Custom credit packages SSO, workspace analytics, configurable teamspaces, priority support, dedicated onboarding, custom team sizes, API

    Credit math (official): same balance → wildly different runtime by model.

    Model Credits/sec Standard (625) Pro (2,250) Max (9,500)
    Gen-4.5 (text-to-video) ~25 ~25 s ~90 s ~380 s
    Gen-4 (image-to-video) ~12 ~52 s ~187 s ~791 s
    Gen-4 Turbo ~5 ~125 s ~450 s ~1,900 s
    Gen-4 Image (1080p) ~8/image — — —

    pricing_type = freemium; starting_price = $0; free_tier = "125 one-time credits (not per month), Gen-4 Turbo only, watermark, 720p, 3 projects, 5GB, commercial allowed — eval/demo only"; pricing_note = "2026-plan-rename basic-to-standard-standard-to-pro-pro-to-max-unlimited-discontinued; max-9500-replaces-unlimited-existing-unlimited-migrate-sep-1-2026; credits-monthly-no-roll-over-except-max-1-month; failed-generations-still-cost-credits; gen-4.5-25-credits-sec-pro-only-90s-mo; api-0-01-credit-retail; annual-20pct-off"

    4. Access Type (v2)

    access_type: web-app, api

    access_display: 🌐 Web App​ (app.runwayml.com / runwayml.com, Chromium-first) · 💻 Desktop​ (heavy GPU rendering, native app available) · 🔌 API​ (REST + Python / Node SDK, $0.01/credit; Enterprise for team/SSO APIs) · 📱 Mobile web limited, no consumer iOS/Android editor​ · (no public free mobile creation app)

    5. Reviews (2026, Split Profile)

    Platform Score Volume Signal
    G2 ~4.3–4.4/5 ~18–80 (still small base) Cinematic quality + feature depth; gripes: credit wall, no roll-over
    Product Hunt / Reddit Mixed 3.5–4.5 Active creator threads “Gen-4 image-to-video is genuinely magical” vs “burned Pro allocation in 3 days”
    Trustpilot ~2–3/5 Billing/refund friction Failed gens refuse refund, credit complaints (same pattern as Cursor/Descript)
    Expert roundups (Top-5 2026) 9.2/10 — “Benchmark suite for studios needing control + APIs, not prompt-only wrappers”

    Praise: best-in-class motion control & camera tooling​ (Motion Brush, Director Mode), character/world consistency unmatched for narrative, Aleph editing loop, model-router breadth, 4K output, professional workflow depth, Discord community 80k+.

    Gripes: credit system is the #1 pain​ — fast consumption, no roll-over (except Max), failed gens not refunded, no live chat support on lower tiers (email/ticket + Discord), 10 s clip ceiling (extend by stitching), occasional morphing on complex prompts, Standard → Pro → Max price jumps steep, desktop-heavy / intermediate skill floor.

    6. Best For / Not For

    ✅ Best For ❌ Not For
    Filmmakers / editors / agencies shipping photoreal or VFX-style shots Casual users wanting “free forever” (125 one-time credits, then $12/mo min)
    Brands needing character/location consistency​ across an ad series Long-form narrative needing 60–180 s single clips (Veo 3 / Kling longer)
    Studios layering generative B-roll into Premiere/DaVinci timelines Beginners who want prompt-only simplicity (Pika/Hedra easier)
    Music videos / short films / concept pre-vis Teams on tight, predictable budgets (credit meter = surprise bills)
    Performance-capture character animation (Act-Two) Pure text-to-app / no-video workflows
    Orgs wanting one subscription, many models​ (Gen-4.5 + Veo + Kling + Seedance) Offline / air-gapped generation (100% cloud)
    API-driven generative pipelines (studios, platforms) Mobile-first creators (no real native editor)

    7. Competitors (Table)

    Tool Lane Starts vs Runway
    Google Veo 3.1​ Foundation text-to-video, native audio, 60 s clips Bundled w/ Google One AI / Vertex AI Longer clips + SynthID + audio; Runway wins on editor/camera control
    OpenAI Sora​ Cinematic text-to-video in ChatGPT ChatGPT Pro/Plus bundle Pushes photorealism + synced audio; no editor suite, no API breadth
    Kling AI (Kuaishou)​ Text/image-to-video, long clips (3 min Pro) ~$5/mo entry More seconds/$; Runway better motion control + ecosystem
    Luma Ray / Dream Machine​ Foundation gen, HDR iteration Freemium Fast iteration; Runway deeper tooling
    Pika​ Casual creator, simple prompt-to-video ~$35/mo Beginner-friendly; Runway more professional
    Hedra​ Character-driven avatars, audio-driven $15–75/mo Character performance; Runway broader scene generation
    HeyGen / Synthesia​ Avatar lip-sync / corporate spokesperson 18+ Talking-head; Runway is scene/cinematic
    Lumen5 / Pictory​ Repurpose text/URL → social clips 25–119 Templated footage-first; Runway is generative
    Descript​ Transcript-edit audio/video $16–50/mo Spoken-word editing; Runway for shot generation
    Premiere / DaVinci / Kapwing​ Timeline NLE 295 / free Frame-accurate editing; Runway for AI generation + light editor

    Closest rivals: Veo 3.1​ (quality + audio + long clips) and Kling AI​ (motion-per-dollar); Runway’s defensible moat = creative control tooling + multi-model router + studio workflow, not raw generation leadership.

     

Do Not Sell or Share My Personal Information Cookie Settings