Abacus.AI sells ChatLLM in two tiers: Basic at $10 per user a month with 20K credits (the first month is $7 with 14K credits) and Pro at $20 a month with 30K credits. Basic is the cheapest way we know to reach Claude Opus 5.5 and GPT 6.1 Sol in one app, but every premium message, image, video and AI Agent task (the agent formerly called DeepAgent) draws from the same credit pool, Basic allows 3 agent tasks a month, and after 75% of your credits the last quarter waits for the final week of your cycle. Abacus doesn't publish credit rates per model, so you see the cost after each message. Zemith Plus is $15.99 a month with 1.5M credits, publishes what research and Creative Studio cost before you run them, and lets you add your own OpenAI, Anthropic and Google keys so those chats use no credits at all.
30K credits; AI Agent and Studio video unrestricted only while credits last; extra credits $10 per 10K
1.5M credits + your own OpenAI, Anthropic or Google key · deep research, Office files, Creative Studio video
Free plan · $0 today · Web, iOS & Android
Pick GPT, Claude, Gemini, Grok or DeepSeek in one chat; models from a provider whose key you added run on that key, everything else runs on credits.
“You can use the last 25% on the last week of your subscription”Abacus.AI billing FAQ, Official answer to "What is this 75% usage message?", Checked 14 Nov 2026
“We constantly add new models, change model rates, and make the product more efficient. This is why you can only see credits after usage.”Abacus.AI billing FAQ, Official answer to "Why don't you have exact credits published per model?", Checked 14 Nov 2026
“Limited Abacus AI Agent access (3 tasks of limited complexity)”Abacus.AI billing FAQ, Official Basic tier limits, Checked 14 Nov 2026
“It's important to note that these will still accrue credits on an ongoing basis, but you will be able to continue using them even after hitting the credit limit.”Abacus.AI billing FAQ, Official answer on unlimited models, Checked 14 Nov 2026
Published US prices, monthly billing, checked 14 November 2026. Abacus credits and Zemith credits are different units and can't be compared one to one.
| Tier | Abacus.AI ChatLLM | Zemith plan | What changes on Zemith |
|---|---|---|---|
Free Try before paying | No free plan on the pricing page; the first Basic month is $7 with 14K credits | $0: 1,100 credits a day on the free models (Qwen, GPT Luna, Gemini Flash, DeepSeek Flash), chat only | You can try Zemith without a card; tools start on Plus |
Entry Everyday chat | Basic $10/user/mo: 20K credits, top LLMs incl. Opus 5.5 on credits, 3 AI Agent tasks, 3 Studio video conversations | Plus $15.99/mo ($179.88/yr): 1.5M credits, paid models, deep research, files, Creative Studio video, plugins | ChatLLM Basic is $5.99 cheaper; Plus has no agent-task or video-conversation cap and adds your own keys |
Top models Claude Opus, GPT 6.1 Sol | Included on Basic, paid from the same 20K credits; the last 25% waits for the final week | Plus with your own OpenAI, Anthropic or Google key, or Pro $29.99/mo on credits | On your key, Opus chats are billed by Anthropic and use no Zemith credits |
Power use Agents, video, long tasks | Pro $20/user/mo: 30K credits, unrestricted AI Agent and Studio while credits last, SuperComputer at 1 credit per 5 minutes | Pro $29.99/mo ($299.88/yr): 3M credits, top models on credits, Cloud tasks and schedules, Max Mode | ChatLLM Pro is cheaper; Zemith Pro adds Cloud agents and keys next to credits |
Out of credits What happens next | Premium models stop until renewal or a purchase: 10K credits cost $10 and roll over; light models keep working | Top-ups from $10 for 800k credits, up to $100 for 10.4M; they never expire | Chats on your own key keep running when credits are low |
Teams Per-seat billing | $10/user/month; each invited user is billed immediately | Extra Plus or Pro seats at the plan price, 1 to 100 seats, bought on the web | Mix Plus and Pro seats in one workspace |
On 14 November 2026 abacus.ai/pricing listed Basic at $10 a month (with a $7 first month) and Pro at $20 a month. The billing FAQ says the subscription is $10 per user per month and Pro is an additional $10 a month. No yearly price or free plan is listed. Abacus.AI Enterprise starts at $5,000.
Basic includes 20K credits a month and Pro 30K (Basic's 20K plus 10K). If you joined on the $7 first-month offer, you get 14K credits that month. Monthly credits reset to the plan amount on your renewal date. Abacus says 20K credits typically cover about 500 images or 25 Abacus AI Agent tasks from a premium service.
Premium models stop until your plan renews or you buy more credits under Profile and Billing, at $10 per 10K credits; purchased credits roll over to the next month. A set of light models (RouteLLM, GPT-6 Luna, Claude Haiku 5.5, Gemini 3.5 Flash Lite and some open models) keeps working, but Abacus says very heavy use after the limit can pause them for the rest of the day. Before that, at 75% of your monthly credits, the last 25% is held for the final week of the cycle.
DeepAgent is now called Abacus AI Agent, and Abacus doesn't publish a fixed credit cost per task: it depends on the task and is shown after use. Basic includes 3 agent tasks of limited complexity a month; Pro removes that cap while you have credits and uses a more powerful version of the agent.
Yes. Teams are billed monthly by team size, and each user you invite is billed immediately at $10 per user per month. Invited members don't pay themselves; the team owner gets the invoice.
You can cancel anytime from the ChatLLM interface (Profile, then Cancel Subscription), and the subscription stays active for the month you paid for. The billing FAQ says Abacus does not offer refunds based on a model's output or performance.
At sticker price, yes: ChatLLM Basic is $10 and Pro $20, against Zemith Plus at $15.99 and Pro at $29.99. Zemith Plus includes 1.5M credits with no agent-task or video-conversation caps, a Free plan, top-ups that never expire, and the option to pay for OpenAI, Anthropic and Google chats with your own key instead of credits.
Founder of Zemith, the all-in-one AI workspace built for people who use GPT, Claude, Gemini, Grok and Creative Studio every day.
I spend my weeks testing model releases, pricing structures and real document, research and creative workflows across Web, iOS and Android.