diff --git a/packages/core/script/generate-venice.ts b/packages/core/script/generate-venice.ts index 5ac11d62..e2562313 100644 --- a/packages/core/script/generate-venice.ts +++ b/packages/core/script/generate-venice.ts @@ -47,6 +47,7 @@ const Pricing = z input: z.object({ usd: z.number(), diem: z.number().optional() }).passthrough(), output: z.object({ usd: z.number(), diem: z.number().optional() }).passthrough(), cache_input: z.object({ usd: z.number(), diem: z.number().optional() }).passthrough().optional(), + cache_write: z.object({ usd: z.number(), diem: z.number().optional() }).passthrough().optional(), }) .passthrough(); @@ -205,6 +206,7 @@ interface MergedModel { input: number; output: number; cache_read?: number; + cache_write?: number; }; limit: { context: number; @@ -276,6 +278,7 @@ function mergeModel( input: spec.pricing.input.usd, output: spec.pricing.output.usd, ...(spec.pricing.cache_input && { cache_read: spec.pricing.cache_input.usd }), + ...(spec.pricing.cache_write && { cache_write: spec.pricing.cache_write.usd }), }; } @@ -348,6 +351,9 @@ function formatToml(model: MergedModel): string { if (model.cost.cache_read !== undefined) { lines.push(`cache_read = ${model.cost.cache_read}`); } + if (model.cost.cache_write !== undefined) { + lines.push(`cache_write = ${model.cost.cache_write}`); + } } // Limit section @@ -409,6 +415,7 @@ function detectChanges( compare("cost.input", existing.cost?.input, merged.cost?.input); compare("cost.output", existing.cost?.output, merged.cost?.output); compare("cost.cache_read", existing.cost?.cache_read, merged.cost?.cache_read); + compare("cost.cache_write", existing.cost?.cache_write, merged.cost?.cache_write); compare("limit.context", existing.limit?.context, merged.limit.context); compare("limit.output", existing.limit?.output, merged.limit.output); compare("modalities.input", existing.modalities?.input, merged.modalities.input); diff --git a/providers/venice/models/claude-opus-45.toml b/providers/venice/models/claude-opus-45.toml index 44ce8a1a..1260301a 100644 --- a/providers/venice/models/claude-opus-45.toml +++ b/providers/venice/models/claude-opus-45.toml @@ -7,13 +7,14 @@ structured_output = true temperature = true knowledge = "2025-03" release_date = "2025-12-06" -last_updated = "2025-12-29" +last_updated = "2026-01-08" open_weights = false [cost] input = 6 output = 30 cache_read = 0.6 +cache_write = 7.5 [limit] context = 202_752 diff --git a/providers/venice/models/zai-org-glm-4.6.toml b/providers/venice/models/zai-org-glm-4.6.toml deleted file mode 100644 index 8c6f0732..00000000 --- a/providers/venice/models/zai-org-glm-4.6.toml +++ /dev/null @@ -1,23 +0,0 @@ -name = "GLM 4.6" -family = "glm-4.6" -attachment = false -reasoning = false -tool_call = true -structured_output = true -temperature = true -knowledge = "2024-04" -release_date = "2025-10-18" -last_updated = "2025-12-29" -open_weights = true - -[cost] -input = 0.85 -output = 2.75 - -[limit] -context = 202_752 -output = 50_688 - -[modalities] -input = ["text"] -output = ["text"]