diff --git a/providers/lanseq/logo.svg b/providers/lanseq/logo.svg new file mode 100644 index 00000000000..dffc0334255 --- /dev/null +++ b/providers/lanseq/logo.svg @@ -0,0 +1,3 @@ + + + diff --git a/providers/lanseq/models/qwen3.8-27b-int4.toml b/providers/lanseq/models/qwen3.8-27b-int4.toml new file mode 100644 index 00000000000..a2e1df9c4a2 --- /dev/null +++ b/providers/lanseq/models/qwen3.8-27b-int4.toml @@ -0,0 +1,26 @@ +# First-party Lanseq deployment documentation: +# https://api.lanseq.cloud/docs +# Pricing: input = $0.25/M tokens, cache_read = $0.045/M tokens, output = $1.99/M tokens. +# Limits: context = 70,000 tokens, max output = 8,192 tokens. +# Reasoning wire control: reasoning_effort = none | low | medium | xhigh. + +base_model = "alibaba/qwen3.8-27b" +name = "Qwen3.8-27B INT4" +description = "Text-only INT4 deployment of Qwen3.8-27B served by Lanseq." +attachment = false + +reasoning_options = [ + { type = "effort", values = ["none", "low", "medium", "xhigh"] }, +] + +[cost] +input = 0.25 +output = 1.99 +cache_read = 0.045 + +[limit] +context = 70_000 +output = 8_192 + +[modalities] +input = ["text"] diff --git a/providers/lanseq/provider.toml b/providers/lanseq/provider.toml new file mode 100644 index 00000000000..f073bad3d64 --- /dev/null +++ b/providers/lanseq/provider.toml @@ -0,0 +1,5 @@ +name = "Lanseq" +npm = "@ai-sdk/openai-compatible" +env = ["LANSEQ_API_KEY"] +api = "https://api.lanseq.cloud/v1" +doc = "https://api.lanseq.cloud/docs"