HeyGen pricing explained: Plans, costs, and the best alternatives

HeyGen pricing looks simple on the plan page and gets complicated on the invoice. What the entry plan buys depends on which avatar model renders the video, how many people share the account, and whether developers call the API. Through 2026, HeyGen retired a plan, renamed its credits, closed unlimited tiers to new buyers, and raised per-minute burn rates without changing list prices.

The plan prices and credit math cover HeyGen's web plans, its API, and its real-time LiveAvatar product. Six alternatives suit teams whose conversations need to happen live.

HeyGen plans and prices in 2026

HeyGen sells five plans, from a free tier to custom Enterprise contracts.

Creator exports at 1080p, while Pro, Business, and Enterprise export at 4K. Business also adds top-up credits and SAML single sign-on (SSO). Some listings still show an older $99 Pro price, which the 2026 price changes explain.

HeyGen credits and burn rates

Credits are HeyGen's single currency, and the avatar model sets the burn rate. After the October 2026 change, per-minute costs run as follows.

• Avatar III: 4 credits with a Video Look, 7 with a Photo Look.
• Avatar IV: 16 credits with a Photo Look, 31 with a Video Look.
• Avatar V: 48 credits.
• Video Agent: 80 credits in standard mode.

Translation is cheaper. Lip-synced translation costs 6 credits a minute in Speed mode and 10 in Precision mode, and translation without lip sync costs 4.

At those rates, Pro's base 1,000 credits cover about 32 minutes of Avatar IV Video Look. Unused monthly credits roll over for one additional billing cycle, and only Business can buy one-off credits, at $0.05 each in 100-credit blocks.

HeyGen API and Enterprise pricing

HeyGen bills its API from a separate prepaid wallet that starts at $5 and requires no subscription. Avatar video bills per second of output, from $0.0167 a second for an Avatar III Digital Twin to $0.0667 a second at premium fidelity, or roughly $1 to $4 per finished minute.

The wallet and the web plan are billed independently. An API key draws on the wallet, while OAuth connections draw on web-plan credits. Enterprise has no published price and is quoted on request, with no maximum video length and 4K export.

HeyGen pricing changes in 2026

HeyGen changed its plans in four separate months of 2026, and many older guides still quote the retired versions.

• January: The $39-per-seat Team plan stopped selling, moving multi-seat buyers to Business at $149 plus $20 a seat.
• February: On February 6, "Generative Credits" became "Premium Credits", and audio dubbing without lip sync stopped consuming credits.
• May: Between May 12 and 21, unlimited plans closed to new customers and Pro dropped from $99 to $49.
• October: Per-minute burn rose at unchanged plan prices, with Avatar IV Video Look going from 20 to 31 credits a minute and Avatar V from 20 to 48.

Legacy subscribers keep their old terms only while their payments stay uninterrupted.

LiveAvatar pricing for HeyGen's real-time avatars

HeyGen's real-time avatars now sit on a separate platform. Interactive Avatar became LiveAvatar, with its own app, documentation, plans, and credits, and the older Interactive Avatar API switched off after a grace period that ended March 31, 2026. LiveAvatar credits are separate from web-plan credits, so a LiveAvatar budget is its own line item.

Paid tiers run from about $19 a month on Starter to about $99 on Essential and $475 on Business. Each streamed minute burns 2 credits in full mode or 1 in avatar-only mode, which works out to roughly $0.08 to $0.10 a minute in avatar-only mode and about double that in full mode.

Billing by streamed minute ties cost to conversation time instead of renders, which is the main difference from HeyGen's credit-based web plans.

HeyGen costs at scale

Spend rises with the avatar model you render, the seats you add, and the credit tier you buy.

Pilot and small-team costs

• On Creator, 600 credits produce very different amounts of footage by model, from 37.5 minutes of Avatar IV Photo Look to about 19 minutes of Avatar IV Video Look or 12.5 minutes of Avatar V.
• Small teams usually land on Business, at $149 for the first seat plus $20 for each additional one. A five-person team pays $$229\ ($149 plus four seats at $20) and shares one pool of 1,500 credits, so each added seat costs money without adding credits.

Scaling production

• Pro scales by credit tier. The base tier is 1,000 credits for $49, the 1,500-credit tier costs $74, and the top tier reaches 100,000 credits for $4,300 a month.
• Heavy use reaches those prices quickly. One creator who needed 20-minute videos estimated about $370 a month on Avatar V and found that too costly at that volume.

Admin and compliance controls by plan

HeyGen gates its admin controls by plan. Business adds SAML SSO, learning management system integrations, and export in the Sharable Content Object Reference Model (SCORM) format. Enterprise adds user provisioning through System for Cross-domain Identity Management (SCIM) and multi-factor authentication (MFA) for team members.

Every new Business seat draws on the same shared 1,500-credit allowance.

HeyGen user reviews

HeyGen reviewers praise avatar realism and lip-sync precision, while billing and support draw complaints. The five main review sites rate it as follows.

• G2: 4.8 from 1,975 reviews.
• Capterra: 4.7 from 320 reviews, with 4.6 for customer service.
• Product Hunt: 4.4 from 72 reviews, with subscription pricing (8) and customer support (7) topping the cons tags.
• Gartner Peer Insights: 4.3 from 28 ratings.
• Trustpilot: 4.2 from 4,940 reviews, the largest sample.

Support experiences vary. On Capterra, one reviewer described a responsive team that resolved an issue quickly, while another said they could not reach a person.

The best HeyGen alternatives

Synthesia, D-ID, and Runway are built mainly for rendered video, though D-ID and Runway now also sell real-time agents. Anam, LemonSlice, and Tavus hold live conversations, and HeyGen sits in both groups through LiveAvatar.

1. Tavus

Tavus is the human computing company, building PALs that see, hear, understand, and respond in real-time conversations. A Personified Application Layer (PAL) is an application a person talks to face to face, one that sees, hears, remembers, and responds. Teams build one on the Conversational Video Interface (CVI), an API that embeds live conversation in a product.

Sparrow-2 governs conversational flow, and Raven-1 fuses the other person's tone, expression, and attention into one signal. The large language model (LLM) layer decides what to say, and Phoenix-4.5 renders the PAL's face, head, and upper body in real time. In preliminary TurnBench results, Sparrow-2 posts 92.4% end-of-turn recall and 97.4% interruption recall.

Product teams configure a PAL for their workflow.

• Knowledge Base: Retrieval-augmented generation (RAG) returns answers in about 30ms, so a PAL coaching a new claims adjuster can quote the exact procedure from an uploaded PDF.
• Function Calling: A PAL can book a calendar follow-up or log a sales note mid-conversation.
• Faces: Custom Faces train from one photo or a one-minute video.

Final Round AI's mock interviews run on Tavus, where Raven-1 picks up shifts in tone and averted gaze and Sparrow-2 gives candidates room to think. More than 100K candidates have completed those interviews, and 75% of users complete two or more.

2. Synthesia

Synthesia builds presenter-led training and internal-communications video from typed scripts. Starter costs $29amonth($18 billed annually) and Pro, formerly Creator, costs $89($64 billed annually), and rates vary by market, so confirm them on Synthesia's pricing page.

• One credit pool: A single balance covers video, dubbing, and surveys, and a finished minute uses about 100 credits.
• AI dubbing: Lip-synced translation into 70+ languages on paid plans at 80 credits a minute, and 140+ on Enterprise.
• Avatar library: 125+ stock avatars on Starter and 180+ on Pro, which adds 5 Personal Avatars.

G2's analysis finds users valuing ease of use and the range of avatars and design kits, and it lists avatar tone and editing flexibility among the cons. Cybernews found the voice sometimes out of step with avatar movement.

3. D-ID

D-ID animates still portraits into lip-synced presenters, and in March 2026 it launched V4 Expressive Visual Agents. Billed annually, Lite costs $4.70 a month, Pro $16, and Advanced $108, with Enterprise priced on request.

• Photo-to-video presenters: Upload a JPEG or PNG portrait and get a talking presenter, exported at 1080p on Trial, Pro, Advanced, and Enterprise.
• V4 Expressive Visual Agents: Real-time, LLM-connected agents launched March 16, 2026.
• API on every plan: API minutes draw from the same balance as the web studio.

G2 reviewers value ease of use and convincing lip sync, and they tag cost and unnatural movement as the repeated complaints.

4. Runway

Runway is a generative video platform built for cinematic production, and its Characters API, launched March 9, 2026, creates real-time video agents from a single image. Standard costs $15amonth($12 billed annually), Pro $35($28), and Max $95($76).

• Gen-4.5: Text and image to video at 12 credits per second, with native audio.
• Runway Characters: Conversational characters built from one reference image, streaming at 24 frames per second in HD.
• Max plan: 9,500 credits a month with one month of rollover, replacing Unlimited for new subscribers.

G2 reviewers praise ease of use and video quality. A Gartner Peer Insights reviewer noted that generation credits can run out quickly during trial and error and that outputs vary from prompt to prompt.

5. Anam

Anam is a developer API for real-time interactive avatars that listen and reply over live video using the LLM you choose.

• Per-minute billing: A free tier includes 30 minutes, and Starter costs $12 a month for 50 minutes, with overage at $0.16 a minute.
• Custom avatars: Unlimited custom avatars on every plan.
• Custom LLM support: Teams can connect their own model and embed personas through the JavaScript SDK.

With only 22 G2 reviews, recurring themes are thin. One reviewer praised the image quality and avatar movement and asked for a faster API.

6. LemonSlice

LemonSlice is a real-time talking-character API with no reviews yet on its G2 seller page.

• Single-image characters: Any photo becomes a real-time avatar that listens, talks, and responds, with no training or fine-tuning.
• Cartoon model: 3D mascots, anime, storybook illustrations, animals, and talking objects from one reference image.
• Open stack: Bring your own LLM or voice model on any tier, or drop a hosted widget into a site with two lines of code.

Starter costs $8 a month and includes 41 minutes, with extra minutes at $0.22.

HeyGen alternatives compared

The table compares delivery mode, live-conversation support, and entry price, with prices as of October 2026.

Credit-based tools bill rendering, so a re-render costs as much as the first pass. Real-time platforms bill minutes of live conversation instead, which makes them easier to compare on cost per conversation.

Choosing a HeyGen plan or a different approach

The right pick depends on what the video has to do once it exists.

• HeyGen fits when the output is presenter-led video made from scripts and monthly volume is predictable. Creator suits a single producer, Pro is the cheaper way to get 4K for one editor, and Business fits teams that need shared seats and SSO.
• HeyGen gets harder to budget when a team re-renders often, mixes avatar models, or adds API and LiveAvatar usage, because each draws on its own credits or wallet.
• A real-time PAL fits when the conversation is the product and the avatar has to listen, notice how the person reacts, remember them, and take actions, as in interview practice, patient intake, or sales coaching. Tavus ships timing, perception, memory, and guardrails under one API, so teams don't wire those pieces together themselves.

A rendered video can be reviewed and fixed before anyone watches it, while a live conversation has to go right in the moment, and that difference usually settles the choice.

The right HeyGen plan comes down to minutes, seats, and credit burn

Teams need to know what their budget will buy before the invoice arrives. With HeyGen, that means accounting for model choice, shared seats, and repeat renders, and then asking whether people need a finished video to watch or a live conversation where they can ask a question and be heard.

Some conversations work better face to face, with someone who listens and answers back. Tavus builds PALs that see, hear, understand, and respond in real time, so a training rehearsal or a customer's question gets an answer in the moment, with real presence.

See it for yourself. Book a demo.