Tingu AI Alternative 2026: A Full Rate Card for Images, and Nothing at All for Chat

TL;DR: Tingu is an all-in-one AI platform whose homepage headline is 'All the AI Models. One Workspace.' and which has begun appearing on AI chat alternative lists as the best all-in-one workspace. Read on 8 September 2026, the site publishes a detailed credit rate card covering roughly a hundred image and video models, from 25 credits for Z Image Turbo to 3,150 credits per five seconds for Kling v3 4K, and publishes no credit rate for a single chat model. The same page says '30+ models included' in one panel and 'One plan. 50+ models' in another. Every price is followed by '+ 18% GST applicable', so the $12 headline is $14.16 at checkout, on a page headed 'Transparent pricing. No hidden fees.' The rate table carries columns for Pro and Scale plans that the pricing cards do not sell, and gives Enterprise a finite 525,000 credits while the Enterprise card promises unlimited generation. Eight models appear twice at different rates. Perspective AI answers the same brief at $14.99/mo with the dollar cost of every action published before it is charged.

Tingu's homepage headline is "All the AI Models. One Workspace." That is close enough to Perspective AI's own pitch that the comparison is worth making carefully, and the honest version of it comes down to one number Tingu does not publish. Tingu has started appearing on AI chat alternative lists, described as the best all-in-one AI workspace. It is a capable product with an unusually detailed rate card. That rate card covers around a hundred image and video models and not one chat model. Everything below was read from tingu.ai on 8 September 2026.

The Rate Card Is Real, and It Stops Before Chat

Credit Tingu with this first, because most of its competitors do not manage it: the site publishes a full per-model, per-generation price table. It is headed "What can you create?" and it lists, for every model, the credit cost and how many generations each pack buys. A sample of the range:

• Z Image Turbo — 25 credits
• Seedream 4.5, GPT Image 1.5, Grok Imagine Image — 50 credits
• Nano Banana, GPT Image 2, Flux 2 Klein — 100 credits
• Nano Banana Pro, Veo 3.1 Lite — 300 credits
• Sora 2, Veo 3.1 (5s) — 600 credits
• Seedance 2.5 — 1,325 credits per 5 seconds
• Sora 2 Pro — 1,800 credits per 5 seconds
• Kling v3 4K and Kling O3 4K — 3,150 credits per 5 seconds

That is good disclosure and it lets you do arithmetic before you pay. Starter's 10,000 credits is 400 Z Image Turbo images, or four five-second Veo 3.1 clips, or three at Kling v3 4K. A buyer can decide on that.

Now look at the filter tabs above the table. They are All, Image, Video. There is no chat tab, and scrolling the table confirms why: there is no chat row. Not one of the language models Tingu advertises has a published credit cost.

This matters because those models are the reason the product is being recommended as a ChatGPT alternative. The provider grid on the same page lists Claude Opus 5, GPT 6 Astra Pro, GPT 5.6 Sol, Grok 4.6, Kimi K3, GLM 5.3, Qwen 3.8 Max and Muse Spark by name. A feature called Chat Arena promises you can "chat with 4 LLMs at once, switch chat models in-between without loosing context". Credits are a single shared balance across every modality. So chat usage spends the same credits as your video clips, at a rate that is published nowhere.

The gap is precise: for images and video, Tingu tells you what a generation costs. For chat, it tells you nothing, on a page whose whole argument is that you should consolidate your chat subscriptions into it.

Two Model Counts, One Page

The comparison panel headed "Stop burning cash on fragmented tools" lists a bullet reading:

"Universal Access — OpenAI, Claude, Gemini, Sora, Veo, Kling, etc. 30+ models included"

Directly beneath it, in larger type, is the heading:

"One plan. 50+ models. No licenses to chase. Just create."

Thirty and fifty, one scroll apart, on the same page on the same day. A feature strip further up also reads "50+ models". Neither figure is qualified, and there is no note explaining that one counts a subset.

The provider grid does not settle it either, because the grid and the site's own changelog disagree. The DeepSeek tile lists one model, an older point release, while the "What's New @ Tingu" feed on the same page announces two newer DeepSeek models by name as now live. The Anthropic tile shows four models, while the changelog announces at least seven distinct Anthropic releases, including two "Fast" variants and a Haiku the tile does not carry at all. Both surfaces are current; they simply are not synchronised, which is a normal thing to happen to a fast-moving catalogue and an abnormal thing to leave on a page selling model access.

"No Hidden Fees" and an 18 Percent Tax Line

The pricing section is headed:

"Transparent pricing. No hidden fees. Pay as you go."

Underneath, each of the three cards prints its price and then, in smaller type, "+ 18% GST applicable". So:

• Starter, advertised $12.00, is $14.16
• Value, advertised $25.00, is $29.50
• Creator, advertised $50.00, is $59.00

Sales tax is legitimate and disclosing it on the card is better than not disclosing it. The problem is the headline immediately above it. "No hidden fees" and an unconditional 18 percent addition are hard to hold at once, and the card does not say which customers the GST applies to, which is the one thing a buyer outside that tax jurisdiction needs to know. A currency toggle offering USD and INR sits below the cards, so the site clearly serves both audiences, and neither is told which of them pays the 18 percent.

There is a second framing issue in the same area. The comparison panel above the cards says "STARTS AT $12 PER MONTH". The cards themselves say the packs have "No expiry on credits", and the section header says "Pay as you go". Those describe a one-time credit purchase, not a monthly subscription. Both readings are on the page, and only one of them supports the "save 95 percent versus $200 a month" arithmetic in the panel.

Columns for Plans You Cannot Buy, and an Unlimited Tier With a Number

The rate table's columns are STARTER, VALUE, CREATOR, PRO, SCALE, ENTERPRISE. The pricing cards offer Starter, Value, Creator and Enterprise. There is no Pro plan and no Scale plan for sale anywhere on the page.

Their sizes can be recovered from the table's own arithmetic. Z Image Turbo costs 25 credits and the table says Starter buys 400 of them, which is 10,000 credits, matching the card exactly. The same division across the remaining columns gives Pro at 100,000 credits and Scale at 260,000. Those are real plans in the pricing model, hidden from the pricing page.

The Enterprise column is the sharper finding. The same division gives Enterprise 525,000 credits, a finite number, consistently across every row in the table. The Enterprise card, meanwhile, sells:

"Unlimited chats, images, videos generation
Unlimited access to ai tools and saas
Unlimited team members"

A tier cannot be both unlimited and 525,000 credits. One of the two surfaces is wrong, and a buyer entering a sales conversation on the strength of the word "unlimited" has no way to know which. This is the most common failure mode in the category: the word "unlimited" almost always has a number behind it, and here the number is printed on the same page.

Eight Models Priced Twice

The table also lists the same model name at two different rates. Filtering out the rows that are simply repeated at an identical price, eight models carry genuinely conflicting figures:

• Kling v2.6 Pro — 250 against 550, per 5s
• Minimax Hailuo 2.3 — 450 against 850, per 5s
• LTX 2 — 100 against 200, per 5s
• Grok Imagine Video v1.5 — 615 against 720, per 5s
• Kling 2.5 Turbo — 500 against 525, per 5s
• Gemini Omni Flash — 750 against 780, per 5s
• Happy Horse (Alibaba) — 1,000 against 1,050, per 5s
• Wan 2.7 — 100 against 750, the second per 5s

The Wan 2.7 pair has an innocent explanation: the provider grid lists Wan 2.7 as both an image and a video model, so the two rows are probably two products sharing a name. That explanation does not cover the others. Kling v2.6 Pro at 250 and at 550 are both marked per five seconds, both in the video set, and differ by 120 percent. The footnote at the bottom of the table says estimates "may vary with resolution, duration, model version, and prompt complexity", which is fair and normal, but it does not explain two fixed credit prices for one name. The practical consequence is that the model name alone does not identify what you will be charged.

The Social Proof Does Not Match the Product

Two smaller things are worth noting because they are the kind of detail that indicates how closely the page is maintained.

A testimonial attributed to a founder reads: "As a solo founder, having access to GPT-4, Claude, and Midjourney all in one place is a game-changer." Midjourney does not appear anywhere in Tingu's model list, provider grid or rate card. GPT-4 is several generations behind everything the changelog announces. The quote describes a product that is not the one being sold.

The numbers in the social-proof strip do not reconcile either. The teams panel claims "70% Avg Savings" and "$1.2M Credits Used", and states that teams save an average of $2,400/year per team member. The comparison panel above claims SAVE 95% against a "$200+ per user / month" baseline. But $2,400 a year is exactly $200 a month, so the claimed saving equals the entire cost being replaced, which would mean the replacement is free. The trust bar below then reports "$2.1M Saved" against the $1.2M of credits used a few hundred pixels earlier. None of these is load-bearing for the product, and that is rather the point: they are the numbers nobody checked.

What to Ask Any All-in-One Platform

Tingu's shape suggests a general test, and it is a short one. When a platform sells many modalities on one shared balance, ask:

1. Is there a published rate for the modality I actually use? A detailed rate card for images tells you nothing about the cost of a conversation if chat has no row.
2. Does the same page give one answer to "how many models"? Two numbers within one scroll means neither is maintained.
3. Is the advertised price the checkout price? Tax lines, annual-billing toggles and "starts at" framings each move the real figure.
4. Does anything described as unlimited have a number elsewhere on the page? Here it does, in the same table, and the number is 525,000.

The pattern generalises well beyond this one vendor. Across the aggregator category, the disclosure is almost always complete for whichever half of the product the vendor built first, and absent for the half it added to widen the pitch. Tingu built a generative-media platform and added chat. The rate card shows exactly where the seam is.

Where Perspective AI Lands

Perspective AI is the chat half done as the main product rather than as an extension. It is $14.99/mo, spanning Anthropic, OpenAI, Google, xAI, DeepSeek, Mistral, Qwen, GLM and Kimi, and every action shows its cost in dollars against your balance before it is charged. That is the number missing from Tingu's page, and it is not a footnote: it is the difference between a subscription you can plan around and one you discover by watching a balance fall. You can switch models mid-conversation without losing the thread, which is the feature Tingu's Chat Arena gestures at. There is no free tier, and stating that here is more useful than letting you find it at a paywall.

If your work is images and video, Tingu has published a rate card you can genuinely budget against, and this article has given it credit for that more than once. If your work is conversation, the honest position is that the price has not been published yet, and a credit balance with no published exchange rate is not a price. Judge every all-in-one claim on the modality you actually use.

FAQ