skip to content
research previewintroducing fx-1 & fx-1 lite. explore the evaluations
artificial hedge. talk to us

models / financial intelligence

the fx series.

frontier depth. economy-optimised capability. financial and mathematical language models from artificial hedge.

explore fx-1explore fx-1 lite
artificial hedgefx series
frontier finance + mathematics
01 / model specificationfrontier / finance + mathematics
frontier model

fx-1

financial context. mathematical depth. a frontier model built around the questions that demand both.

research preview / waitlist

depth for demanding work.

our State-of-the-Art frontier model for demanding financial and mathematical work. built for deep reasoning across complex analytical tasks.

total parameters
3.4T
active parameters / token
~128B
model focus
State-of-the-Art frontier finance and mathematics
model usage subscriptions
frontier economy · frontier extended · warp+
API rates / 1m tokens
$10 input$30 output
join fx-1 waitlist view subscription tiers
02 / model specificationeconomy / finance + mathematics
economy model

fx-1 lite

the financial and mathematical focus of the fx series, optimised for an economy model offering.

research preview / waitlist

capability, economically.

a highly capable model optimised for financial and mathematical work. exceptional reasoning capability built for economy pricing.

total parameters
2.4T
active parameters / token
~49B
model focus
high-capability finance and mathematics at economy pricing
model usage subscriptions
economy · extended · warp+
API rates / 1m tokens
$5 input$15 output
join fx-1 lite waitlist view subscription tiers

comparative specifications

two perspectives.
one clear comparison.

inspect model size or API rates. total parameters, active parameters and token prices each describe a different part of the offering.

comparison lens

total parameters

trillions / T
fx-13.4T
fx-1 lite2.4T

the full parameter count of each model.

active per token

billions / B · approximate
fx-1~128B
fx-1 lite~49B

~128B for fx-1; ~49B for fx-1 lite.

each chart has its own scale. parameter counts describe model size; they are not measures of performance.

fx-1 / evaluation record

read the results.
keep the context.

reported fx-1 results span finance, mathematical reasoning, research, documents and workflows. each evaluation measures its own task.

explore the evaluation record
Finance Agent v263.5scoreFinanceBench100scoreGPQA Diamond97.3%percentageOmniDoc Bench98score

company-reported fx-1 results. no fx-1 lite benchmark scores are presented. evaluation configurations and methodology will accompany the technical release.

your model enquiry

start with
your priorities.

choose the model focus you want to discuss. your selection carries through to the waitlist enquiry.

select a model focus

selected for your enquiry

fx-1

state-of-the-art frontier finance and mathematics

total parameters
3.4T
active / token
~128B
join fx-1 waitlist your selection reflects your stated priority. model suitability depends on evaluation within your workflow.

application directory

the work
around the model.

explore the financial workflows guiding our research.

the next step

bring a demanding
question.

discuss the model, the financial task and the evidence you would use to evaluate it.

discuss model access

explore artificial hedge

local site index↑↓ navigate ↵ open