feat(meta): Add pricing and capabilities configuration

[mayankjain-panw]

Summary

Adds pricing and model capabilities configuration for Meta AI provider (Muse Spark models).

Changes

• pricing/meta.json — Token pricing for all 3 models + default zero-cost baseline + web_search billing
• general/meta.json — Model parameters and capability declarations

Pricing Source (from Meta docs)

Model	Input ($/1M)	Output ($/1M)	Cached ($/1M)	Web Search ($/1K)
muse-spark-1.1	$1.25	$4.25	$0.15	$2.50
muse-spark-1.2	$1.25	$4.25	$0.15	$2.50
muse-spark-1.2-contributor	$0.10	$0.20	$0.002	$2.50

Conversion: price_value = dollars_per_million / 10_000 (verified against x-ai and anthropic entries).

Web search: 0.0025 = $2.50/1000 per-query rate (same unit convention as x-ai web_search field).

Capabilities (general/meta.json)

• Context window: 1,048,576 tokens
• Primary type: chat
• Supported: tools, image, video
• Message roles: system, developer, user, assistant, tool
• stop parameter intentionally omitted (Meta rejects on reasoning models)
• n parameter omitted (only n=1 supported)

Pricing Structure (pricing/meta.json)

• default entry: zero-cost baseline for unrecognized models
• Per-model entries with request_token, response_token, cache_read_input_token, cache_write_input_token ($0), additional_units.web_search
• calculate.request: sum of (input_tokens * rate) + (cache_read_input_tokens * cache_rate)
• calculate.response: output_tokens * rate

Validation

Both JSON files pass JSON.parse() validation. 112 insertions, 2 new files.

Related PRs

• Gateway: Portkey-AI/gateway-enterprise-node#1907
• Albus: Portkey-AI/albus#2477
• Protego: Portkey-AI/protego#97
• Luna: Portkey-AI/luna#1963

[Copilot]

Pull request overview

Adds Meta provider configuration to the models repository by introducing token pricing and model capability/parameter metadata for Muse Spark models, enabling downstream systems to price requests and understand supported features.

Changes:

• Added pricing/meta.json with per-model token and cached-token pricing (plus a default baseline pricing config).
• Added general/meta.json with provider/model parameter definitions and declared supported modalities (tools/image/video).

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.

File	Description
pricing/meta.json	Introduces Meta Muse Spark pricing configuration and a default cost calculation formula.
general/meta.json	Declares Meta Muse Spark model parameters and capabilities (tools/image/video).

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated no new comments.

Suppressed comments (1)

pricing/meta.json:29

• calculate.request references cache_read_input_tokens, but the pricing schema/examples elsewhere in this repo use the metric name cache_read_tokens (and cache_write_tokens). As written, cached-token usage will likely be excluded from cost calculation, and cache-write costs won’t be counted if they become non-zero in the future.

            {
              "operation": "multiply",
              "operands": [
                { "value": "cache_read_input_tokens" },
                { "value": "rates.cache_read_input_token" }
              ]