{ "updated": "2026-10-15", "currency": "GBP", "contact": "agent@riskmandate.ai", "assumptions": { "fx_usd_per_gbp": 1.30, "model_prices_usd_per_mtok": { "source": "Anthropic first-party prices, cached 6 Oct 2026", "fable-5.1": {"input": 10.00, "output": 50.00, "cache_read": 0.25}, "opus-5.5": {"input": 4.00, "output": 20.00, "cache_read": 0.20}, "sonnet-5.5": {"input": 2.00, "output": 10.00, "cache_read": 0.20}, "haiku-5.5": {"input": 0.10, "output": 0.50, "cache_read": null, "note": "prompts up to 100K tokens"} }, "ec2_usd_per_hour": { "source": "eu-west-2 (London) on-demand, from third-party aggregators (doit.com, devzero.io); to be confirmed on AWS's price list", "small": {"instance": "t3.medium", "vcpu": 2, "memory_gib": 4, "usd_per_hour": 0.0472}, "medium": {"instance": "m7i.xlarge", "vcpu": 4, "memory_gib": 16, "usd_per_hour": 0.2331}, "gpu": {"instance": "g5.xlarge", "vcpu": 4, "memory_gib": 16, "gpu": "1x NVIDIA A10G", "usd_per_hour": 1.277} }, "notes": [ "All prices in GBP, excluding VAT. No payment is taken on the site: the order is emailed to agent@riskmandate.ai and we reply with a quote.", "Token mix (editable): 75% cache reads, 20% fresh input, 5% output. That blends to $1.05 per million tokens on Sonnet 5.5, $1.95 on Opus 5.5, $0.05 on Haiku 5.5 (Haiku cache reads assumed at $0.01) and $4.69 on Fable 5.1.", "Token budgets are Sonnet-equivalent: 1M on Sonnet 5.5 = 0.5M on Opus 5.5 = 10M on Haiku 5.5 (half the cost-equivalent 20M, because Haiku's cache price is not listed). Fable 5.1 on request.", "We add 25% on top of the workers' tokens for the queen and the host session (not yet measured).", "Measured on our own runs (15 Oct 2026): 75k to 151k tokens per worker, 325k to 541k per run with 3 to 6 workers, 17 to 21 minutes per run; studio run 1 used 629,857 tokens across 7 builders.", "Runs are not automated end to end yet: a person on our side starts, watches and hands over every run. Prices cover that person's time at an assumed £75 an hour.", "Margin target 40%, floor 35%. The arithmetic behind every price is in out/ruv-sgit/r5-market/economist/economics.md." ] }, "groups": [ { "id": "runs", "title": "Task runs", "summary": "A Ruflo queen and a team of agent workers do one task on our shared infrastructure, inside an sgit vault, under an Agent Behaviour Policy with no internet access. You get the vault back.", "who": "Anyone with a well-defined task and the materials for it: consultants, teams trying ruv.sgit, people who want the audit trail as much as the result.", "map": "runs", "items": [ { "id": "run-s", "name": "Small run", "price": 10, "unit": "per run", "summary": "A queen and up to 5 workers for a focused task.", "details": [ "1 queen + up to 5 workers in parallel", "Up to 1M tokens on Sonnet 5.5, or 0.5M on Opus 5.5, or 10M on Haiku 5.5", "Stops at 15 minutes of wall-clock time", "Materials up to 50 MB, uploaded by you", "No network access: the run's behaviour policy allows no egress", "You get the run's sgit vault: every brief, message, delivery, decision and the run ledger entry, as a zip or published with a read key", "If a cap is reached the run stops; the vault holds everything done until then" ], "options": [ {"id": "model", "label": "Model", "choices": [ {"id": "sonnet", "label": "Sonnet 5.5 (1M tokens)", "delta": 0}, {"id": "opus", "label": "Opus 5.5 (0.5M tokens)", "delta": 0}, {"id": "haiku", "label": "Haiku 5.5 (10M tokens)", "delta": 0}]}, {"id": "key", "label": "Model key", "choices": [ {"id": "ours", "label": "Our key (tokens included)", "delta": 0}, {"id": "byok", "label": "Bring your own key (tokens billed to your account)", "delta": -1}]}, {"id": "delivery", "label": "Delivery", "choices": [ {"id": "zip", "label": "Vault as a zip", "delta": 0}, {"id": "published", "label": "Vault published with a read key (cannot be undone)", "delta": 0}]} ], "max_qty": 50 }, { "id": "run-m", "name": "Team run", "price": 20, "unit": "per run", "summary": "A queen and up to 10 workers: the size of our own site-building runs.", "details": [ "1 queen + up to 10 workers in parallel", "Up to 3M tokens on Sonnet 5.5, or 1.5M on Opus 5.5, or 30M on Haiku 5.5", "Stops at 25 minutes of wall-clock time", "Materials up to 100 MB, uploaded by you", "No network access: the run's behaviour policy allows no egress", "You get the run's sgit vault: every brief, message, delivery, decision and the run ledger entry, as a zip or published with a read key", "If a cap is reached the run stops; the vault holds everything done until then" ], "options": [ {"id": "model", "label": "Model", "choices": [ {"id": "sonnet", "label": "Sonnet 5.5 (3M tokens)", "delta": 0}, {"id": "opus", "label": "Opus 5.5 (1.5M tokens)", "delta": 0}, {"id": "haiku", "label": "Haiku 5.5 (30M tokens)", "delta": 0}]}, {"id": "key", "label": "Model key", "choices": [ {"id": "ours", "label": "Our key (tokens included)", "delta": 0}, {"id": "byok", "label": "Bring your own key (tokens billed to your account)", "delta": -3}]}, {"id": "delivery", "label": "Delivery", "choices": [ {"id": "zip", "label": "Vault as a zip", "delta": 0}, {"id": "published", "label": "Vault published with a read key (cannot be undone)", "delta": 0}]} ], "max_qty": 50 }, { "id": "run-l", "name": "Large run", "price": 50, "unit": "per run", "summary": "A queen and up to 20 workers for a big task.", "details": [ "1 queen + up to 20 workers in parallel", "Up to 10M tokens on Sonnet 5.5, or 5M on Opus 5.5, or 100M on Haiku 5.5", "Stops at 45 minutes of wall-clock time", "Materials up to 100 MB, uploaded by you", "No network access: the run's behaviour policy allows no egress", "You get the run's sgit vault: every brief, message, delivery, decision and the run ledger entry, as a zip or published with a read key", "If a cap is reached the run stops; the vault holds everything done until then" ], "options": [ {"id": "model", "label": "Model", "choices": [ {"id": "sonnet", "label": "Sonnet 5.5 (10M tokens)", "delta": 0}, {"id": "opus", "label": "Opus 5.5 (5M tokens)", "delta": 0}, {"id": "haiku", "label": "Haiku 5.5 (100M tokens)", "delta": 0}]}, {"id": "key", "label": "Model key", "choices": [ {"id": "ours", "label": "Our key (tokens included)", "delta": 0}, {"id": "byok", "label": "Bring your own key (tokens billed to your account)", "delta": -10}]}, {"id": "delivery", "label": "Delivery", "choices": [ {"id": "zip", "label": "Vault as a zip", "delta": 0}, {"id": "published", "label": "Vault published with a read key (cannot be undone)", "delta": 0}]} ], "max_qty": 20 } ] }, { "id": "compute", "title": "Dedicated compute", "summary": "An AWS EC2 instance in London started for your task and stopped after it (SG Compute). The vault is created inside that environment, so nothing is shared with other customers.", "who": "Tasks that need their own machine, more memory, a GPU or a longer time than a shared run allows.", "map": "compute", "items": [ { "id": "compute-small", "name": "Small instance", "price": 21, "unit": "per 8-hour block", "summary": "t3.medium: 2 vCPU, 4 GiB memory.", "details": [ "AWS EC2 t3.medium in eu-west-2 (London): 2 vCPU, 4 GiB", "Price = £20 platform fee + twice the EC2 cost ($0.0472 an hour)", "Started for your task, stopped when it ends or after 8 hours", "The sgit vault is created inside the instance", "Add a task run for the tokens (Task runs above)" ], "max_qty": 30 }, { "id": "compute-medium", "name": "Medium instance", "price": 23, "unit": "per 8-hour block", "summary": "m7i.xlarge: 4 vCPU, 16 GiB memory.", "details": [ "AWS EC2 m7i.xlarge in eu-west-2 (London): 4 vCPU, 16 GiB", "Price = £20 platform fee + twice the EC2 cost ($0.2331 an hour)", "Started for your task, stopped when it ends or after 8 hours", "The sgit vault is created inside the instance", "Add a task run for the tokens (Task runs above)" ], "max_qty": 30 }, { "id": "compute-gpu", "name": "GPU instance", "price": 36, "unit": "per 8-hour block", "summary": "g5.xlarge: 4 vCPU, 16 GiB memory, 1 NVIDIA A10G.", "details": [ "AWS EC2 g5.xlarge in eu-west-2 (London): 4 vCPU, 16 GiB, 1 NVIDIA A10G GPU", "Price = £20 platform fee + twice the EC2 cost ($1.277 an hour)", "Started for your task, stopped when it ends or after 8 hours", "The sgit vault is created inside the instance", "Add a task run for the tokens (Task runs above)" ], "max_qty": 30 } ] }, { "id": "repo", "title": "Maintained repository", "summary": "The full ruv.sgit setup in a private repository of your own, kept up to date by us. Runs on Claude Code, on AWS or anywhere else you run agents.", "who": "Consultants, companies and internal teams who want to run ruv.sgit themselves without maintaining it.", "map": "repo", "items": [ { "id": "repo-individual", "name": "Individual", "price": 149, "unit": "per month", "summary": "One private repository for one person, kept current.", "details": [ "The full ruv.sgit setup in a private repository", "New versions as we release them", "Setup help by email", "Up to 1 hour of our time a month; more as expert time", "Model tokens and cloud costs are on your own accounts" ], "max_qty": 12 }, { "id": "repo-team", "name": "Team", "price": 499, "unit": "per month", "summary": "One repository for a team, with your branding and a better UI.", "details": [ "Everything in Individual", "Custom branding and the latest UI", "A setup call, and help running it on Claude Code or AWS", "Up to 3.5 hours of our time a month; more as expert time", "Model tokens and cloud costs are on your own accounts" ], "max_qty": 12 }, { "id": "repo-white-label", "name": "White-label ready", "price": 1500, "unit": "per month", "summary": "Ready to offer to your own clients under your name.", "details": [ "Everything in Team", "White-label ready: your name, your domain, no mention of ours unless you want it", "Priority on new versions and fixes", "Up to 10 hours of our time a month; more as expert time", "rUv's projects stay attributed as their licences require (MIT, by rUv / ruvnet)", "Model tokens and cloud costs are on your own accounts" ], "max_qty": 12 } ] }, { "id": "expert", "title": "Expert time", "summary": "Time with the people who built ruv.sgit, to scope tasks, steer runs and read the results with you. Mix it with runs: ten runs and two hours to steer them works well.", "who": "Anyone buying runs who wants help shaping them, and teams adopting ruv.sgit.", "map": "expert", "items": [ { "id": "expert-hours", "name": "A couple of hours", "price": 500, "unit": "per session", "summary": "About 2 hours: scope a task, steer a few runs, review what came back.", "details": [ "About 2 hours of expert time, remote", "Good with a pack of runs: we shape the briefs and read the vaults with you", "Runs are not automated end to end yet: a person on our side starts, watches and hands over each run" ], "max_qty": 10 }, { "id": "expert-day", "name": "A day", "price": 1500, "unit": "per day", "summary": "A working day: design a run pipeline, a behaviour policy, or a pilot.", "details": [ "One working day of expert time, remote", "Design and run a pilot with you, including its Agent Behaviour Policy", "Runs are not automated end to end yet: a person on our side starts, watches and hands over each run" ], "max_qty": 5 }, { "id": "expert-days", "name": "Several days", "price": 5000, "unit": "per engagement", "summary": "Four working days: set ruv.sgit up in your environment and hand it over.", "details": [ "Four working days of expert time", "Set up ruv.sgit in your environment (Claude Code, AWS or your own), with your policies and your first runs", "Runs are not automated end to end yet: a person on our side starts, watches and hands over each run" ], "max_qty": 3 } ] } ] }