visual-ai-assertions 0.19.0 → 0.20.0

package/README.md

@@ -579,6 +579,7 @@ Any [OpenRouter](https://openrouter.ai/models) model slug (always `vendor/model`
 
 | Model          | Model ID                    | Input $/MTok | Output $/MTok | Notes                                 |
 | -------------- | --------------------------- | ------------ | ------------- | ------------------------------------- |
+| Muse Spark 1.3 | `meta/muse-spark-1.3`       | $1.25        | $4.25         | Meta flagship, 1M context; gated¹     |
 | Grok 4.6       | `x-ai/grok-4.6`             | $2           | $6            | Newest xAI flagship, 500K context     |
 | Grok 4.5       | `x-ai/grok-4.5`             | $2           | $6            | Prior xAI flagship, 500K context      |
 | Kimi K3        | `moonshotai/kimi-k3`        | $3           | $15           | Moonshot flagship, 1M context         |
@@ -587,6 +588,10 @@ Any [OpenRouter](https://openrouter.ai/models) model slug (always `vendor/model`
 | Qwen3.7 Plus   | `qwen/qwen3.7-plus`         | $0.32        | $1.28         | Cost-effective, GUI/screen-reading    |
 | Qwen3.6 Flash  | `qwen/qwen3.6-flash`        | $0.19        | $1.13         | **Default** — cheap flash vision tier |
 
+¹ Muse Spark 1.3 is age-gated by OpenRouter: calls return HTTP 403 (`VisualAIAuthError`) until the account completes the 18+ confirmation at [openrouter.ai/settings/preferences](https://openrouter.ai/settings/preferences). It also reasons by default — expect several hundred reasoning tokens per call even with no `reasoningEffort` set.
+
+Meta also publishes `meta/muse-spark-1.3-contributor`, the same model at $0.10 / $0.20 per MTok — about 12x cheaper — because Meta uses everything submitted through it for product improvement. It has **no named constant** (`Model.OpenRouter` does not expose it) and never appears by default anywhere in this library, so using it takes a deliberate, explicit choice: pass the slug directly as a plain string, `visualAI({ model: "meta/muse-spark-1.3-contributor" })`. Any OpenRouter slug works this way — see the note above the table — and cost tracking works correctly once you opt in. OpenRouter itself blocks it with HTTP 404 (`paid-model-training-violation-by-account`) until the account's privacy settings allow training endpoints, at [openrouter.ai/settings/privacy](https://openrouter.ai/settings/privacy). Only use it if sending your screenshots to Meta for training is a trade you've deliberately made.
+
 `qwen/qwen3.7-max` and the DeepSeek V4 family (`deepseek/deepseek-v4-pro`, `deepseek/deepseek-v4-flash`, and dated variants such as `deepseek/deepseek-v4-pro-0813`) are not listed because they accept no image input on OpenRouter.
 
 ## License

package/dist/index.cjs

@@ -126,6 +126,7 @@ var Model = {
    * recognizes them. All listed models accept image input.
    */
   OpenRouter: {
+    MUSE_SPARK_1_3: "meta/muse-spark-1.3",
     GROK_4_6: "x-ai/grok-4.6",
     GROK_4_5: "x-ai/grok-4.5",
     KIMI_K3: "moonshotai/kimi-k3",
@@ -1263,8 +1264,26 @@ var PRICING_TABLE = {
     inputPricePerToken: 0.5 / PER_MILLION,
     outputPricePerToken: 3 / PER_MILLION
   },
-  // OpenRouter passes through upstream per-model pricing (verified 2026-07-22
+  // OpenRouter passes through upstream per-model pricing (verified 2026-09-03
   // against https://openrouter.ai/api/v1/models).
+  // Meta's own listed rates ($1.25 / $4.25, cached input $0.15) match
+  // OpenRouter's pass-through exactly. Cached input is not modelled here:
+  // `calculateCost` applies no cache discount on any provider.
+  [`${Provider.OPENROUTER}:${Model.OpenRouter.MUSE_SPARK_1_3}`]: {
+    inputPricePerToken: 1.25 / PER_MILLION,
+    outputPricePerToken: 4.25 / PER_MILLION
+  },
+  // Muse Spark 1.3's data-sharing tier: same model, ~12x cheaper, because
+  // Meta trains on everything submitted through it. Deliberately keyed by the
+  // literal slug rather than a `Model.OpenRouter` entry — it must never be
+  // reachable via autocomplete or default selection. Pass the string yourself
+  // (`model: "meta/muse-spark-1.3-contributor"`) to opt in; cost is still
+  // tracked correctly once you do. Cached input is $0.002/MTok, not modelled
+  // (no provider gets a cache discount here).
+  [`${Provider.OPENROUTER}:meta/muse-spark-1.3-contributor`]: {
+    inputPricePerToken: 0.1 / PER_MILLION,
+    outputPricePerToken: 0.2 / PER_MILLION
+  },
   [`${Provider.OPENROUTER}:${Model.OpenRouter.GROK_4_6}`]: {
     inputPricePerToken: 2 / PER_MILLION,
     outputPricePerToken: 6 / PER_MILLION