Platform plans for teams, rate cards for training, datasets, and compute. Rent what we've built, bring your own, or commission something new — our open research stays free for everyone.
Plan rates
prices in USD · cancel anytime
Researcher
For individuals exploring LegionEdge protocols and open research.
$0
forever
TeamPopular
For teams building production AI systems on LegionEdge.
$49
per seat / month
Enterprise
For organizations with custom security & compliance needs.
Custom
annual contract
Team trial
14 days free — no card required
Seat changes
Prorated on the next invoice
Open research
Free for everyone, on every plan
Estimate
prices in USD
Plan
Seats
Billing
Team bills per active seat and prorates changes; annual billing takes two months off. Open research needs no plan at all.
Compare plans
Researcher
Team
Enterprise
Platform
Sandbox environments
Seats
AI orchestration tooling
Early access to new protocols
Products
Nokuva
Tavoc
Foltrac
Research
Open datasets & protocol specs
Private model fine-tuning
Custom synthetic data services
Support & security
Support
SSO / SAML
Custom SLAs
Onboarding
Two axes set the price of every run: the compute it uses, metered per GPU-hour, and the parameters you need, billed per billion. Fine-tune, shrink a model you already have, or scope a custom program.
Training rates
compute used + params needed
Up to 8B params
$3,500
per B params / run
8B – 34B params
$3,000
per B params / run
34B – 70B params
$2,500
per B params / run
70B+ params
Custom
scoped with the lab
Compute used
Metered per GPU-hour at Compute rates — no markup
Eval runs
Included on every run
Weights & checkpoints
Yours, on every engagement
Minimum contract
$40,000 — dataset & training engagements, scoped
Other engagements
quoted after a scoping call
Distill & optimize
A smaller student, quantized and tuned to serve on your hardware.
$2,000
per B params, student
Frontier custom
Pretraining programs with a dedicated researcher on the project.
Custom
scoped with the lab
Estimate
compute + params · H100 on-demand
Params needed
Training tokens
Method
A run bills as compute used at the Compute-tab GPU rates plus the per-params service fee; custom programs are quoted after a scoping call. The weights are yours either way. Dataset and training engagements start at a $40,000 minimum contract.
High-quality training data three ways, all priced per billion params of the model you train: license the corpora our pipeline keeps refreshing, run your own data through it, or commission a corpus built for your domain.
License our corpora at a rate that scales with the params of the model you train.
Run your own data through the generation and validation pipeline, priced by the model it feeds.
A synthetic corpus built for your domain and your model scale, delivered with its harness.
LE-Code
Code generation & repair
1.2T tokens
Monthly
$600 / B / mo
LE-Interface
UI/UX interaction traces
240M episodes
Monthly
$500 / B / mo
LE-Reasoning
Verified reasoning chains
85M chains
Quarterly
$750 / B / mo
Open research releases
Published with our papers
Various
—
Free forever
Dataset pricing scales with model size — a 7B model pays 7× the per-B rate, a 70B model 70×. Rented corpora carry a non-exclusive term license; anything we process or build for you is yours, including the generation harness. Dataset and training engagements start at a $40,000 minimum contract.
The same GPU fleet behind our research, priced per GPU-hour with no instance minimums. Reserve for a year and the rate drops about 30%.
GPU rates
priced per GPU-hour · billed hourly
Accelerator
VRAM
On-demand
Reserved (1 yr)
B200
180 GB
$6.90
$4.85
H100
80 GB
$3.95
$2.75
A100
80 GB
$2.40
$1.70
L40S
48 GB
$1.15
$0.85
NVMe scratch storage
$0.08 / GB-month
Egress
Free up to 10 TB / month, then $0.02 / GB
InfiniBand interconnect
Included on multi-node jobs
Estimate
30-day month
Accelerator
GPUs
Usage
Commitment
Estimates assume a 30-day month and steady daily usage. Reserved rates require a 12-month commitment; spot capacity is available on request.
Everything we publish stays open, whatever your plan. Rent the lab's time when you want it pointed at your problem.
Everything the lab publishes — papers, datasets, protocol specs, and harnesses.
Our researchers working across the whole board with your team.
Point the lab at your problem; the findings publish openly either way.
Open by default
what publishing openly means here
Papers
CC BY 4.0
Research datasets
Open data license
Protocol specs
Apache 2.0
Eval harnesses
MIT
Sponsorship buys attention, not conclusions — sponsored findings publish openly whether or not they flatter the sponsor.
Yes — every Team workspace starts with a 14-day free trial, no card required. When it ends you can downgrade to Researcher without losing your data.
You're billed monthly for active seats. Add or remove people any time; changes prorate automatically on your next invoice.
The Startup Program includes up to $25k in product credits over 12 months for eligible early-stage teams — often more valuable than a discount.
Nothing. Datasets, protocol specs, and published research are open to everyone regardless of plan — that's the point of publishing them.
You do — weights, checkpoints, and eval results. Training engagements are work for hire; we keep the methods, you keep the model.
By the GPU-hour, with no instance minimums — on-demand meters exactly the hours you use. Reserved capacity bills monthly against a 12-month commitment at roughly 30% below on-demand.
Tell us about your workload and we'll tell you honestly — including when the free tier is all you need.
