ResolveAI routes every step to the best fit for quality, latency, and cost.
Frontier quality, priced for everyday workloads.
For incidents and paging alerts, when every minute to a root cause counts.
Median, at Standard rates
* By default, incidents run on Resolve Pro Fast. Typical costs are medians; your cost depends on the work, and every run shows its cost live.
Frontier quality, priced for everyday workloads.
For incidents and paging alerts, when every minute to a root cause counts.
Lower price per token, and more incidents solved.
50%
less than Claude Fable 5.2 and GPT-6.2 Astra
$4 input and $20 output, against $8 and $40. The same price as Claude Opus 5.5, with better quality.
25%
less than the fast modes
$12 input and $60 output, against $16 and $80.
| Per million tokens | Input | Output | Cache write | Cache read |
|---|---|---|---|---|
| $4 | $20 | $5 | $0.20 | |
| Claude Opus 5.5 | $4 | $20 | $5 | $0.20 |
| Claude Fable 5.2 | $8 | $40 | $10 | $0.20 |
| GPT-6.2 Astra | $8 | $40 | $10 | $0.40 |
| Per million tokens | Input | Output | Cache write | Cache read |
|---|---|---|---|---|
| $12 | $60 | $15 | $0.60 | |
| Claude Fable 5.2 fast | $16 | $80 | $20 | $0.40 |
| GPT-6.2 Astra fast | $16 | $80 | $20 | $0.80 |
US dollars per million tokens, list prices, excluding tax, November 2026. Cache write is the shortest retention each provider offers.
2×
faster to the coding agents' best pass rate
Answers in minutes, updates as the investigation continues.
Pass rate: share of runs whose RCA scores at least 0.8 against a verified reference, a threshold we calibrated with expert SREs.
Inside the Resolve agentic system →Pay as you go at published rates, or get pricing built for your teams.
Pay-as-you-go, for teams getting started.
Includes:
For many teams that need more control and governance.
Everything in Standard, plus:
Run ResolveAI in your own environment, deployed to your requirements.
Everything in Enterprise, plus:
| Feature | Standard | Enterprise | Custom |
|---|---|---|---|
| Products | |||
| Alert investigations | |||
| Issues | |||
| Incidents | |||
| Chat | |||
| Scheduled tasks | |||
| Live dashboards | |||
| Shareable reports | |||
| Users | Up to 100 | Unlimited | Unlimited |
| Integrations | Up to 5 | Unlimited | Unlimited |
| Live cost per investigation and spend limits | |||
| Early access to new products and models | — | ||
| Security and compliance | |||
| SAML/OIDC SSO | — | ||
| SCIM provisioning | — | ||
| Role-based access control | — | ||
| Audit logs | — | ||
| Custom data retention, PII redaction | — | ||
| Spend and reporting | |||
| Team spend limits | — | ||
| Chargeback by team | — | ||
| Usage analytics | — | ||
| Alert hygiene report | — | ||
| Team reliability reports | — | ||
| Support | |||
| Priority support | — | ||
| Forward-deployed engineers | — | — | |
| Deployment | |||
| Hosted in | ResolveAI's cloud | ResolveAI's cloud | Single-tenant or your VPC |
| Customer-managed encryption keys | — | — | |
| Data residency | — | — | |
| Bring your own model | — | — | |
| Usage and billing | |||
| Resolve Units (1 RU = $1) | Pay as you go | Annual commitment | Annual commitment |
| Payment | Credit card | Invoice | Invoice |
Use your existing AWS agreement
AWS MarketplaceAlso available as plugins for Slack, Cursor, Anthropic, and OpenAI.
1 RU = $1 of usage, prepaid.
Investigates an alert or incident, answers a chat or runs a task.
Everything the AI reads and writes, at your model's rate. You see each investigation's cost live.
Only usage draws from your balance.
About three quarters of a word. ResolveAI reads logs, metrics, code and alerts, and writes its reasoning and answers, all in tokens. A quick chat uses few tokens; a hard incident uses many.
A Resolve Unit (RU) is how you prepay for usage. One RU equals $1 of usage at the published token rates. You buy Resolve Units up front, and the work ResolveAI does draws them down.
Every investigation, chat and task is priced by the tokens it uses, at the rate of the model family you chose for it. Input, output and cached tokens each have their own rate, listed in the token rates above.
Usage draws down your Resolve Unit balance as the work happens. On Standard you pay as you go. On Enterprise and Custom you commit to an annual amount of Resolve Units, and usage draws from that commitment.
Yes. Set a spend limit on every run, and a monthly cap for your whole organization and for each team. A run that hits its limit pauses for a person to approve more.
Only the limits you set. Standard includes up to 100 users and 5 integrations. Enterprise and Custom have no limit on users or integrations.
Resolve Pro for everyday workloads, and Resolve Pro Fast for incidents and paging alerts, where every minute counts. Resolve SRE-1, trained by ResolveAI Labs for production work, is coming soon. You choose a model family for each alert rule, chat or task, and Custom plans can bring their own model.
Background work such as alert investigations and scheduled tasks is usage like any other, priced by the tokens it uses. Your spend limits and team caps apply to it, and every run shows its cost.
Standard is billed monthly by credit card, with no commitment. Enterprise and Custom are invoiced against an annual commitment. You can also buy through AWS Marketplace with your existing AWS agreement.
Prices in US dollars, excluding tax. Rates are list prices per million tokens. When our published rates fall, you get the lower rate.