feat(standardcompute): add Standard Compute provider (flat-rate OpenAI-compatible router)

providers/standardcompute/models/standardcompute.toml

@@ -0,0 +1,32 @@
+# tool_call and structured_output verified against the live endpoint
+# (https://api.stdcmpt.com/v1/chat/completions) on 2026-07-17.
+# release_date reflects the public launch month (March 2026) as stated by the
+# operator; the provider does not publish a launch-date page.
+# limit.context and limit.output are conservative placeholders (the router
+# fans out to frontier backends whose true limits vary per request); chosen
+# to match the floor convention used by other auto-router entries.
+# attachment/reasoning are set false pending verification, not as claims of
+# absence.
+name = "Standard Compute (auto-route)"
+description = "Flat-rate auto-router: one model id that routes each request to a suitable frontier model, unlimited tokens at a fixed monthly price"
+family = "auto"
+release_date = "2026-03-01"
+last_updated = "2026-07-17"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = true
+structured_output = true
+open_weights = false
+
+[cost]
+input = 0.00
+output = 0.00
+
+[limit]
+context = 128_000
+output = 16_384
+
+[modalities]
+input = ["text"]
+output = ["text"]

providers/standardcompute/provider.toml

@@ -0,0 +1,12 @@
+# Flat-rate unlimited gateway: per-token cost is $0 because pricing is
+# subscription-based — tiers from $9 to $399/mo that differ by execution
+# speed, not usage caps. Pricing page: https://standardcompute.com/pricing
+# (accessed 2026-07-18). The single "standardcompute" model id routes each
+# request to a suitable frontier model. OpenAI-compatible chat completions;
+# tool calls and structured output verified against the live endpoint
+# 2026-07-17. Setup docs: https://standardcompute.com/integrations
+name = "Standard Compute"
+env = ["STANDARDCOMPUTE_API_KEY"]
+npm = "@ai-sdk/openai-compatible"
+doc = "https://standardcompute.com/integrations"
+api = "https://api.stdcmpt.com/v1"