Models & Pricing

stepIQ is model-agnostic. Each step can use a different model, and runtime prices are calculated from the same constants used by the API and worker.

Available models

Anthropic

Claude Haiku 3.5

claude-3-5-haiku-20241022 • Max tokens: 8,192

$3.13 / 1M in • $15.63 / 1M out

Anthropic

Claude Opus 4.6

claude-opus-4-6 • Max tokens: 8,192

$62.50 / 1M in • $312.50 / 1M out

Anthropic

Claude Sonnet 4

claude-sonnet-4-20250514 • Max tokens: 8,192

$37.50 / 1M in • $187.50 / 1M out

Google

Gemini 2.5 Flash

gemini-2.5-flash • Max tokens: 65,536

$3.75 / 1M in • $31.25 / 1M out

Google

Gemini 2.5 Pro

gemini-2.5-pro • Max tokens: 65,536

$15.63 / 1M in • $125.00 / 1M out

Mistral

Mistral Large Latest

mistral-large-latest • Max tokens: 32,000

$25.00 / 1M in • $75.00 / 1M out

Mistral

Mistral Small Latest

mistral-small-latest • Max tokens: 32,000

$2.50 / 1M in • $7.50 / 1M out

OpenAI

GPT-4o

gpt-4o • Max tokens: 16,384

$31.25 / 1M in • $125.00 / 1M out

OpenAI

GPT-4o Mini

gpt-4o-mini • Max tokens: 16,384

$1.88 / 1M in • $7.50 / 1M out

OpenAI

GPT-5.2

gpt-5.2 • Max tokens: 128,000

$21.88 / 1M in • $175.00 / 1M out

OpenAI

GPT-5.2 Instant

gpt-5.2-chat-latest • Max tokens: 128,000

$21.88 / 1M in • $175.00 / 1M out

OpenAI

GPT-5.3-Codex

gpt-5.3-codex • Max tokens: 128,000

$21.88 / 1M in • $175.00 / 1M out

z.ai

GLM-4.6

glm-4.6 • Max tokens: 128,000

$1.00 / 1M in • $2.50 / 1M out

z.ai

GLM-5

glm-5 • Max tokens: 128,000

$0.25 / 1M in • $1.38 / 1M out

These are the same marked-up prices returned by GET /api/models (base provider price + 25% markup).

Credit system

Credits are tracked per account plan. Token-to-credit conversion is implemented as 1 credit = 1,000 tokens for legacy/BYOK-style deduction paths.

Plan	Included credits	Overage rate per credit
Free	100	N/A
Starter	2,000	€0.010
Pro	8,000	€0.008
Enterprise	Custom	N/A

Cost estimation

Current implementation computes cost from measured token usage at execution time. There is no dedicated pre-run estimation endpoint yet.

Per step:

base_cost_cents =
  (input_tokens / 1_000_000) * input_cost_per_million +
  (output_tokens / 1_000_000) * output_cost_per_million

step_cost_cents = ceil(base_cost_cents * (1 + 25 / 100))

Run totals are the sum of step costs. Credits are then deducted by funding mode:

app_credits (Starter/Pro with platform keys): ceil(total_cost_cents / overage_per_credit_cents)
legacy (e.g. free): ceil(total_tokens / 1,000)
byok_required: 0 credits deducted by platform

run-cost-example.txt

Run: weekly-blog-generator
Funding mode: app_credits (starter)

Total tokens: 42,600
Total cost: 62 cents
Credits deducted: ceil(62 / 1) = 62 credits

If the same run used legacy mode:
credits deducted: ceil(42,600 / 1,000) = 43 credits