Runway Review 2026: The AI Video Pioneer That Quietly Renamed Itself and Stopped Claiming #1
Runway was one of the first companies to make AI video generation feel real, and it's still shipping: a new flagship model, a video editor, a color-grading model, motion capture, and real-time character agents, all launched or updated in the last twelve months. But two things changed underneath that momentum by the time we checked. First, runwayml.com now 308-redirects to runway.com, and the company's own homepage no longer leads with "AI video generator" -- it leads with "Building Real-World Intelligence," a pivot toward general world models for robotics and interactive interfaces. Second, the #1-on-Artificial-Analysis claim Runway made when it launched Gen-4.5 in December 2025 doesn't hold up against Artificial Analysis' own live leaderboard today -- Gen-4.5 isn't on it at all, and an entirely different set of providers now leads. Neither of those facts erases what Runway does well. But a genuinely rough 1.1/5 Trustpilot record (344 reviews) means this is a review worth reading before you buy credits, not after.
| Ownership | Independent, founder-led since 2018 |
| Free tier | Real -- 125 one-time credits |
| Cheapest paid plan | $12/mo (Standard, annual) |
| Valuation (Feb 2026, reported) | ~$5.3B, Series E led by General Atlantic |
| Trustpilot record | 1.1/5, 344 reviews -- prompt/refund complaints |
Runway still ships faster than almost anyone in this category. The reason to pause is how it treats the credits you already paid for when the output isn't what you asked for.
"What surprised me most about checking Runway in September 2026 wasn't any single feature -- it's that the company barely talks about 'AI video generation' anymore. Its own homepage now opens with 'Building Real-World Intelligence,' the old runwayml.com domain 308-redirects to runway.com, and two of its newest research posts (GWM Worlds 2, Solaris) are about robots and interactive interfaces, not video clips. Underneath that pivot, the actual video product kept shipping hard: Gen-4.5 in December, Aleph 2.0 for editing in May, Ruby for color grading and native DaVinci Resolve/Adobe plugins by September. But when Runway launched Gen-4.5, it said the model ranked #1 on Artificial Analysis' leaderboard -- and when I checked that same leaderboard live on the day I verified this review, Gen-4.5 wasn't on it, at all, and the field is now led by names (Alibaba's Wan, ByteDance's Seedance, MiniMax) most buyers have never heard of. Meanwhile Runway's Trustpilot record sits at 1.1 out of 5 across 344 reviews, with a complaint pattern that isn't about the company being a scam -- it's specifically about credits being spent on generations that don't follow the prompt, with no refund path for 'completed but wrong' output."
None of that means Runway is bad software -- Amazon Prime, BBC Studios, Lionsgate and Under Armour are all real, named users, and the platform's breadth (generation, editing, color, motion capture, character agents) is genuinely wide. It does mean I'd test on the Free or Standard tier before committing real budget to Pro or Max, read the credit/refund terms closely, and not assume a launch-day "#1 model" claim is still true by the time you're paying for it.
Three platforms now, not one video tool: Runway Creative, Runway Dev, and -- new for 2026 -- Runway Robotics.
Runway started as a browser-based video-editing/generation tool and has grown into three distinct platforms, per its own current homepage. Runway Creative is the cloud suite most people mean when they say "Runway": Gen-4.5 (flagship video generation), Aleph 2.0 (video editing, now inside "Edit Studio"), Act-Two (motion/performance capture), Ruby (SDR-to-HDR color grading), Runway Characters (real-time conversational video agents from one image), and Agent 2.0 (an AI agent for building multi-platform marketing content). Runway Dev is the API platform (still living at dev.runwayml.com) offering "frontier models and the best from other labs, behind one key," plus two separate MCP surfaces for coding agents. Runway Robotics is the newest and most unexpected: a simulation/policy-inference toolkit built on GWM-1, Runway's General World Model, aimed at training and evaluating robot behavior in simulated environments before deploying it in the real world.
Worth flagging up front: Runway has not been acquired -- it's still run by its 2018 co-founders (Cristobal Valenzuela, Alejandro Matamala, Anastasis Germanidis) -- but its own public identity has moved noticeably since Gen-4.5 launched. The domain change and the "Real-World Intelligence" homepage rewrite are Runway's own choices, not evidence of a takeover.
A reported $5.3B valuation and real studio customers -- next to a benchmark claim that didn't survive a live re-check.
Wikipedia-sourced figure; Runway's own Series E post discloses the raise but not a valuation number.
Series E, closed Feb 10, 2026, led by General Atlantic.
Per Wikipedia -- notably small next to the reported valuation.
NYC, by 3 co-founders who met at NYU's ITP program.
| Date | Event | Result |
|---|---|---|
| 2018 | Valenzuela, Matamala & Germanidis found Runway | $2M seed |
| 2022-12 | Series C | $50M at $500M valuation |
| 2023-06 | Series C extension, led by Google/Nvidia/Salesforce | $141M more at $1.5B valuation |
| 2025-04-03 | Series D, led by General Atlantic (official post) | $300M+ raised; ~$3B+ valuation per Wikipedia |
| 2025-12-01/04 | Gen-4.5 launched; claimed #1 on Artificial Analysis (Elo 1,247) | General paid-plan availability Dec 4, 2025 |
| 2026-02-10 | Series E, led by General Atlantic again (official post) | $315M raised; ~$5.3B reported valuation |
| 2026-09-29 | This review's verification date | runwayml.com -> runway.com redirect confirmed; Gen-4.5 absent from Artificial Analysis' live leaderboard |
Trustpilot is genuinely rough -- and JAVIS could not reach a second rating source to check whether that's the whole picture.
1.1 out of 5, with 93% one-star reviews. The dominant complaint: the tool "does not listen to your prompts, only grabs your money," with users reporting characters changing appearance between clips and objects rendering incorrectly, needing multiple paid re-generations to approach something usable.
Multiple reviewers describe the credit system as "exploitative," and support is quoted as denying refunds because "credits are only auto-refunded for system-detected failures, not for completed-but-wrong output" -- i.e. if the generation finishes but doesn't match the prompt, there's reportedly no refund path.
A friendlier, editorial take: praises Runway's help content specifically ("great help content -- check out the Academy") and its range of purpose-built models, but names a real con: "steep learning curve" -- confirms the same Free/Standard ($12/mo annual) pricing JAVIS verified directly.
Six tiers, from a real (if small) free plan to custom Enterprise -- on the same credit-based billing model driving the complaints above.
Standard
625 credits/mo ($12/mo billed annually). 20GB storage.
- All AI video & image models, no watermark
- 4K upscaling, up to 3 projects
Pro
2,250 credits/mo ($28/mo annual). 100GB storage.
- 1 Brand Kit, 1 custom voice clone
- Runway MCP, unlimited 4K upscaling
Max
9,500 credits/mo, 1-month rollover ($76/mo annual).
- 3 Brand Kits, 3 voice clones
- HDR/ProRes studio-grade output
Team
6,900 credits/seat/mo, pooled ($55/seat annual). 2-9 seats.
- 1TB storage, API access
- Shared Agent Skills, Figma/Dropbox/Notion connectors
A real Free plan exists below Standard: $0/mo, 125 one-time credits (not renewing monthly), 5GB storage, limited model selection. Above Team: Enterprise is custom-priced and is the only tier that unlocks model-weights licensing, SSO, and dedicated security/audit-report access (see Security). Credits are shared across models but spend at different rates per model -- see API/Economics for the exact per-second costs.
Your answer depends on how much you trust a launch-day benchmark claim versus wanting to test it yourself first.
Free plan -- 125 one-time credits is enough for a handful of short test generations; read the Trustpilot pattern above first so you know what "failure" does and doesn't get refunded.
Standard, $12/mo annual -- the real entry point once Free's one-time credits are gone.
Pro, $35/mo -- the tier most individual professionals should evaluate.
Max, $95/mo -- 9,500 credits/mo with rollover and studio-grade formats.
Team, $69/seat/mo -- pooled credits, shared projects, API access baked in.
Enterprise, custom -- these are gated features, not available on any self-serve plan including Max or Team.
Check Alternatives below -- Runway's own Dec 2025 #1 claim isn't corroborated on Artificial Analysis' current leaderboard, where Kling 3.0 currently outranks both Veo 3.1 and Sora 2.
A genuinely capable flagship model -- launched with a benchmark claim that a live re-check couldn't confirm ten months later.
Gen-4.5 is Runway's flagship text-to-video and image-to-video model, reaching general paid-plan availability on Dec 4, 2025 (the launch research post itself is dated Dec 1, 2025). Runway's own announcement claims: realistic object weight, momentum and force; liquids that "flow with proper dynamics"; handling of complex, multi-element scene compositions; "believable collisions and natural movement"; and character expression with "nuanced emotions, natural gestures and lifelike facial detail," spanning photorealistic through stylized-animation looks. Runway states Gen-4.5 "maintains the speed and efficiency of Gen-4" and is priced the same across subscription plans, running on Nvidia Hopper and Blackwell GPUs. At launch, Runway cited an Artificial Analysis Text-to-Video Arena score of 1,247 Elo, ahead of Google Veo 3 (1,226), Kling 2.5 (1,225) and OpenAI Sora 2 Pro (1,206) -- a genuine #1 claim at that specific moment, on that specific leaderboard.
| Claim | At Gen-4.5 launch (Dec 2025, per Runway) | Live re-check (Sep 29, 2026, per Artificial Analysis) |
|---|---|---|
| Gen-4.5 rank, Text-to-Video Arena | #1, Elo 1,247 | Not listed on the current leaderboard at all |
| Google Veo (3 at the time, 3.1 now) | #2, Elo 1,226 | 14th, Elo 1,088 (Text-to-Video) |
| Kling (2.5 at the time, 3.0 now) | #3, Elo 1,225 | 12th, Elo 1,095 (Text-to-Video) |
| OpenAI Sora (2 Pro at the time, 2 now) | #4, Elo 1,206 | 17th, Elo 1,083 (Text-to-Video) |
| Current #1 (Sep 2026) | N/A | Gemini Omni Flash (Google), Elo 1,233 |
Published by the official "Runway" YouTube channel (youtube.com/@RunwayML), checked 2026-09-29.
- Gen-4.5 is priced the same as Gen-4 across every plan -- Runway didn't use the new model as an excuse to raise prices.
- The feature claims (multi-element scenes, physical dynamics, character expression range) are consistent with what a genuinely capable current-generation video model should do, and match the general direction the whole category has moved.
- Runway continued shipping real, dated updates around Gen-4.5 all year (Aleph 2.0, Act-Two, Ruby, Characters) rather than going quiet after one launch.
- Don't take a vendor's own launch-day leaderboard citation as still true months later -- this review's live re-check found Gen-4.5 absent from Artificial Analysis' current leaderboard entirely.
- The field has been reshaped by providers most Western buyers have never evaluated (Alibaba's Wan, ByteDance's Seedance, MiniMax's H3) -- worth a side-by-side test against Gen-4.5 before assuming brand recognition equals current quality.
Prompt or footage in, produced clip out -- the review step is where the Trustpilot complaints actually originate.
The human-review step is exactly where the Trustpilot complaints above originate: if a generation doesn't follow the prompt but technically "completes," re-running it spends more credits with (per reviewer reports) no refund path for the first attempt. Separately, Runway announced a "unified pricing" change on Sep 18, 2026 that merges its web-app and Dev-API credit pools into one fungible balance with centralized usage tracking -- a genuine simplification of how credits are managed across products, though it does not appear to change the specific refund-for-bad-output policy Trustpilot reviewers describe.
The part of Runway that's grown the most unexpectedly: real-time characters, a content-scaling agent, and simulated interactive worlds.
Beyond straight video generation, 2026 has been the year Runway pushed hardest into agent- and world-shaped products. Runway Characters (Jul 19, 2026) turns a single image into "a real-time conversational video agent," with synchronized facial animation, lip-sync and head movement at 24fps in HD. Agent 2.0 (Jun 25, 2026) is a content-production agent aimed at marketing teams -- building campaign assets, running data analysis, and scaling content across multiple platforms from one conversation, extended by Timeline-in-Agent (Jul 8, 2026) and Custom Agent Skills (Jul 6, 2026) that let an agent build and share reusable skills across a workspace. Separately, and more experimentally, GWM Worlds 2 (Sep 3, 2026) "turns high-fidelity video and audio generation into real-time interactive simulation" -- you define a world and steer it with text commands while it generates 720p video at 24fps live.
- Runway Characters and Agent 2.0 are shipped, dated, real products -- not research-only demos -- aimed at concrete marketing/content use cases.
- GWM Worlds 2 and Solaris (interface generation) represent a genuinely distinct strategic bet most video-generation competitors aren't publicly making.
- A real-time conversational video agent built from someone's likeness (Characters) is a materially higher-sensitivity surface than a static generated clip -- review consent and usage terms before deploying one representing a real person.
- The strategic pivot toward world models/robotics is real R&D investment, but it also means some of Runway's newest launches (GWM-1, Worlds 2, Solaris) aren't things most current paying video customers will ever touch.
Native plugins inside DaVinci Resolve and Adobe, plus two separate MCP servers for coding agents.
As of the Sep 2026 changelog, Runway ships a native DaVinci Resolve plugin (Sep 23, 2026) and native Adobe Premiere Pro / After Effects plugins (Sep 8, 2026), both confirmed directly on runway.com/changelog, letting editors call Runway's models without leaving their existing edit. Separately, Runway runs two genuinely distinct MCP (Model Context Protocol) surfaces, confirmed on its own developer docs: a generation MCP (mcp.runwayml.com, the "Runway MCP" feature listed on the Pro plan and above, newly listed on Cursor's Marketplace as of Sep 23, 2026) for creating media directly inside a chat interface, and a separate Dev-portal MCP server (dev.runwayml.com/mcp) that lets a coding agent look up API tasks, manage model routers, and read docs -- explicitly documented to work with Cursor, Claude Desktop, Claude Code (via a one-line claude mcp add command), and Codex.
| Integration | What it does | Where it's documented |
|---|---|---|
| DaVinci Resolve plugin | Bring Runway's generative tools directly into an existing Resolve edit | runway.com/changelog (official, Sep 23, 2026) |
| Adobe Premiere Pro / After Effects plugins | Same, inside Adobe's edit surfaces | runway.com/changelog (official, Sep 8, 2026) |
| Generation MCP (mcp.runwayml.com) | Generate video/image/audio directly inside a chat client; now on Cursor's Marketplace | docs.dev.runwayml.com (official) |
| Dev-portal MCP (dev.runwayml.com/mcp) | Coding agent looks up API tasks, manages model routers, reads docs | docs.dev.runwayml.com/guides/mcp/ (official) -- confirms Claude Desktop & Claude Code support |
One credit meter, different burn rates per model -- $0.01 per credit on the Dev API.
| What you're generating | What it costs |
|---|---|
| Evaluating on the Free plan | $0 -- 125 one-time credits (does not renew monthly) |
| Generating on Standard/Pro/Max | $15-$95/mo -- 625 to 9,500 credits/mo included |
| Gen-4.5 video (API) | 12 credits/second ($0.12/second at $0.01/credit) |
| Gen-4 Turbo video (API) | 5 credits/second ($0.05/second) |
| Act-Two motion capture (API) | 5 credits/second ($0.05/second) |
| Aleph 2 editing (API) | 28 credits/second, 56-credit minimum per generation ($0.28/sec, $0.56 min) |
| Gen-4 Image (API) | 5 credits/720p image, 8 credits/1080p image |
| Gen-4 Image Turbo / Muse Image (API) | 2 credits / 1 credit per image respectively |
| Image upscaling (Magnific Precision v2) | 25 credits/image (150 credits if output exceeds 4096px) |
| Video upscaling (Magnific Video Upscaler Creative) | $0.007-$0.012 per output frame, by resolution |
| Rolling out to a company | Enterprise, custom -- model-weights licensing, SSO, dedicated security review bundled into a contract |
Real certifications, a real layered safety model -- with the deepest controls gated to Enterprise, same as most of this category.
| Control | What's actually in place |
|---|---|
| Certifications | SOC 2 Type II (annual, AICPA-audited), ISO/IEC 27001:2022 (per runway.com/data-security) |
| Regulatory alignment | GDPR (EU), CCPA/CPRA (California), NIST-framework elements |
| Customer data / training | "We don't train on your data" (Enterprise page); production content kept isolated from testing environments |
| Content provenance | C2PA participation, per runway.com/safety |
| Moderation model | Prevent (data filtering, pre-launch red-teaming) / Detect (AI classifiers, human review of reports) / Enforce (automatic suspension of repeat/unlawful-content violators) |
| Enterprise-only controls | SSO, model-weights licensing, granular access management, direct Trust Center audit-report access, dedicated security review |
Runway's certifications and safety architecture are real and independently checkable claims, not just marketing copy -- SOC 2 Type II and ISO 27001:2022 are specific, auditable standards. What's worth separating clearly: the Enterprise page's "we don't train on your data" statement applies to Enterprise/API customer content specifically, and the deepest governance controls (SSO, audit-report access, model-weights licensing) are not available on Free through Team -- the same self-serve/Enterprise split JAVIS has found across most of this category.
Best for people who want one platform spanning generation, editing and color -- not people who need a guaranteed refund for a bad generation.
| Situation | Why Runway fits | Likely plan |
|---|---|---|
| Solo creator testing AI video for the first time | Real Free plan (125 credits) to try Gen-4.5 before paying anything | Free, then Standard |
| Video editor wanting AI tools inside an existing NLE | Native DaVinci Resolve and Adobe Premiere/After Effects plugins | Pro or Max |
| Small marketing/content team scaling output together | Team plan's pooled credits, shared projects, Agent 2.0 for multi-platform content | Team ($69/seat) |
| Developer building on Claude/Cursor who wants Runway generation as a tool | Generation MCP (mcp.runwayml.com) and a separate Dev-portal MCP for account management | Pro+ for MCP access, or Runway Dev API directly |
| Studio/brand needing licensing, SSO, or dedicated security review | Enterprise -- real named customers include Amazon Prime, BBC Studios, Lionsgate, Under Armour | Enterprise, custom |
| Buyer who wants the objectively highest-ranked model today, brand aside | Not clearly Runway's current strength -- see the live leaderboard re-check in Core Feature | Compare against Kling AI or the current AA leaders first |
Five real gaps, not manufactured ones.
The most consistent theme across 344 Trustpilot reviews: generations that don't follow the prompt but "complete" successfully are, per reviewer reports, not eligible for a credit refund -- only system-detected failures reportedly are.
Gen-4.5 does not appear at all on Artificial Analysis' live Text-to-Video or Image-to-Video leaderboards as of this review's verification date -- a genuine, checkable gap between the launch-day marketing claim and the current state of the objective benchmark.
G2, Capterra, GetApp, TrustRadius, SaasWorthy and Reddit were all inaccessible during this research -- an honest evidentiary gap, not a claim that Trustpilot is definitely the whole picture.
SSO, model-weights licensing, and dedicated audit-report access all require custom-priced Enterprise -- there's no self-serve path, even on the $69/seat Team plan.
Aleph 2 (video editing) spends credits over 5x faster per second than Gen-4 Turbo -- a plan chosen based on the pricing page's "X seconds of Gen-4.5" framing can run out much faster once editing work enters the mix.
Runway isn't the only, or currently the highest-ranked, AI video option -- and one competitor prices identically to it.
| If you mostly need... | Compare Runway with... |
|---|---|
| The exact same headline pricing tiers with a different generation engine | Pika Labs -- Creator ($35/mo) and Fancy ($95/mo) price-match Runway's Pro and Max exactly |
| Video generation bundled into tools you already pay for | Google Veo 3.1 -- via Gemini app, Flow, Vertex AI, Workspace |
| You already pay for ChatGPT and want video without a new subscription | OpenAI Sora 2 / Sora 2 Pro |
| Currently higher-ranked raw quality than Veo/Sora on an independent benchmark | Kling AI 3.0 (Kuaishou) -- ranks 12th on Artificial Analysis vs. Veo's 14th and Sora's 17th, as of this check |
| A broader multi-modal creative workspace, not just video | Luma AI |
| The current objective #1 on Artificial Analysis, brand-name recognition aside | Google's Gemini Omni Flash, Alibaba's Wan 3.0, or MiniMax's H3 -- all outrank every household-name video model on the live leaderboard right now |
| Vendor | Comparable self-serve mid-tier price |
|---|---|
| Runway Standard | $12/mo annual (625 credits) |
| Pika Starter | $8/mo annual (900 credits) |
| Runway Pro | $28/mo annual (2,250 credits) |
| Pika Creator | $28/mo annual (3,150 credits) -- identical headline price to Runway Pro |
| Luma Plus | $25/mo annual (10,000 credits -- a different, larger credit unit, not directly comparable 1:1) |
| Runway Max | $76/mo annual (9,500 credits) |
| Pika Fancy | $76/mo annual (8,550+ credits) -- identical headline price to Runway Max |
No affiliate relationship shapes this review.
JAVIS has not established a publisher affiliate program with Runway. Every CTA in this article points to the official product.
Go to the official page.
Open RunwayRead the alternative that matches your real question.
Choose the next articleRunway makes the most sense next to the AI voice tool it naturally pairs with, and the Claude/MCP ecosystem its Dev platform plugs into.
Runway is JAVIS's first AI Video flagship review, so there's no direct same-category review yet. The most genuinely relevant next reads are the AI voice platform that pairs naturally with Runway's video output, the Claude/MCP ecosystem Runway's own Dev-portal MCP server explicitly supports, and the automation platform many Runway users script their generation pipelines through.
Each card image is a real, already-verified official asset -- 4 are the same real assets used on the target article itself (ElevenLabs' own cover image, Claude's own hero asset, a real Claude MCP illustration, a Mintlify-generated Claude Code OG image, n8n's own official brand image, and Lovable's own official funding-announcement cover image); none were guessed.
Questions people are actually searching right now
Did Runway change its name or website?
Its public domain did: runwayml.com now issues a live HTTP 308 permanent redirect to runway.com, and the company's own homepage now leads with "Building Real-World Intelligence" rather than an AI-video-generator pitch. The legacy "RunwayML" name survives as the developer/API subdomain (dev.runwayml.com) and on Runway's still-unmigrated Trustpilot listing. This is not an acquisition -- Runway remains independently run by its 2018 co-founders.
How much does Runway actually cost?
A real (if small) Free plan exists: 125 one-time credits. Self-serve paid plans run $15-$95/month (Standard/Pro/Max), or 20% less with annual billing, plus a Team plan at $69/seat/month (2-9 seats) and custom-priced Enterprise. All figures verified live on runway.com/pricing and runway.com/team-plan, Sep 29, 2026.
Is Gen-4.5 really the best AI video model?
It was a genuine, corroborated #1 claim on Artificial Analysis' Text-to-Video Arena at its Dec 2025 launch (Elo 1,247). Checking that same leaderboard live on Sep 29, 2026 found no Runway Gen-4.5 entry at all -- the field has been reshaped by providers like Google's Gemini Omni Flash, Alibaba's Wan 3.0, and MiniMax's H3. This doesn't prove Gen-4.5 got worse; it means a ~10-month-old vendor benchmark claim shouldn't be assumed still current.
What is Runway worth?
Runway's own Series D (Apr 2025) and Series E (Feb 2026, $315M, led by General Atlantic) announcements disclose the funding amount but not a valuation figure. Wikipedia's sourced summary reports ~$3B+ after Series D and a ~$5.3B valuation after Series E -- third-party-reported numbers, not Runway's own disclosed figure.
Is Runway good, based on reviews?
Runway's Trustpilot rating is genuinely poor: 1.1 out of 5 across 344 reviews (93% one-star), with a specific, repeated complaint about credits spent on generations that don't follow the prompt and no refund path for "completed but wrong" output. JAVIS could not independently verify a G2, Capterra, or Reddit rating in this research to check whether a friendlier professional-user population exists, as is the case for some competitors -- a real evidentiary gap, disclosed rather than filled with a guess.
Does Runway work with Claude or MCP?
Yes, in two distinct ways, per Runway's own developer docs: a generation MCP (mcp.runwayml.com, the "Runway MCP" feature on Pro+ plans) for creating media inside a chat client, and a separate Dev-portal MCP server (dev.runwayml.com/mcp) explicitly documented to work with Claude Desktop and Claude Code for managing your API account from a coding agent.
How does Runway compare to Pika, Veo, Sora and Kling?
Pika prices its Creator and Fancy tiers identically to Runway's Pro and Max ($35/$95 monthly). Google Veo 3.1 and OpenAI Sora 2 are typically bundled into platforms you may already pay for (Gemini/Vertex, ChatGPT) rather than sold as a standalone subscription. On Artificial Analysis' live leaderboard as of this check, Kling AI 3.0 currently ranks ahead of both Veo 3.1 and Sora 2 -- and all three currently rank below several less consumer-familiar providers (Alibaba, ByteDance, MiniMax, Google's Gemini-branded models).
Where this came from, and when it was checked.
Domain rebrand (runwayml.com -> runway.com): direct check, confirmed 308 permanent redirect, Sep 29, 2026.
Pricing (Free/Standard/Pro/Max/Team/Enterprise): runway.com/pricing and runway.com/team-plan, Sep 29, 2026.
Homepage positioning, three-platform structure, safety/security claims: runway.com, runway.com/safety, runway.com/data-security, runway.com/enterprise, official, checked Sep 29, 2026.
Company background, founders, HQ, funding timeline: en.wikipedia.org/wiki/Runway_(company), cross-checked against Runway's own official Series D and Series E announcement blog posts for the funding-amount and lead-investor facts.
Series D ($300M+, Apr 3, 2025) and Series E ($315M, Feb 10, 2026): runway.com/news/runway-series-d-funding and runway.com/news/runway-series-e-funding, official Runway posts (valuation figures sourced separately from Wikipedia, not from these posts).
Gen-4.5 launch claims and benchmark citation: runway.com/research/introducing-runway-gen-4.5, official, plus runway.com/changelog for the Dec 4, 2025 general-availability date.
Live leaderboard re-check (the article's central finding): artificialanalysis.ai/video/leaderboard/text-to-video and artificialanalysis.ai/video/leaderboard/image-to-video, both checked Sep 29, 2026, full ranked lists captured and checked for any "Runway" entry.
Product timeline (Aleph 2.0, Act-Two, Ruby, Characters, Agent 2.0, GWM-1, GWM Worlds 2, Solaris): runway.com/changelog and each product's own research/news post (introducing-runway-aleph, introducing-act-one, introducing-runway-gwm-1, introducing-gwm-worlds-2, news/introducing-aleph-2-and-edit-studio, news/introducing-runway-characters, news/research/introducing-solaris), all official.
Integrations (DaVinci Resolve, Adobe, MCP): runway.com/changelog and docs.dev.runwayml.com/guides/mcp/, official.
API/credit pricing: docs.dev.runwayml.com/guides/pricing/, official.
Trustpilot rating (1.1/5, 344 reviews): successful direct read of trustpilot.com/review/runwayml.com, Sep 29, 2026.
G2/Capterra/GetApp/TrustRadius/SaasWorthy/Reddit: all attempted and found inaccessible during this research -- disclosed rather than omitted silently.
Independent editorial review: zapier.com/blog/best-ai-video-generator/, checked, corroborates pricing and adds independent pros/cons.
Competitor pricing (Pika, Luma): pika.art/pricing and lumalabs.ai/dream-machine/pricing, official, checked.
Competitor model status (Veo, Sora, Kling): deepmind.google/models/veo/ (official) plus the same Artificial Analysis live leaderboards used for the Runway re-check.
Official video: youtube.com/watch?v=AwKSrJFvdps, authorship confirmed on YouTube (author "Runway", channel @RunwayML).
