feat(azure): add GPT-5.5 model metadata

## Summary

Adds GPT-5.5 metadata for:

- Azure
- Azure Cognitive Services

The Azure Cognitive Services entry mirrors the existing local convention of full TOML model definitions.

## Sources

- Microsoft Learn lists `gpt-5.5` for Azure OpenAI / Microsoft Foundry with version `2026-04-24`, `1,050,000` context, `922,000` input, `128,000` output, structured outputs, tools, image input, and December 2025 training data.
- Microsoft’s Azure GPT-5.5 announcement lists pricing at `$5.00` input, `$0.50` cached input, and `$30.00` output per 1M tokens.
- Azure Responses API docs list PDF input support for vision-capable models and include `gpt-5.5` version `2026-04-24`.

## Notes

This intentionally does not add `gpt-5.5-pro`, since Azure Learn currently lists `gpt-5.5` but not `gpt-5.5-pro` in the Azure model catalog.

This also intentionally omits OpenAI-specific `context_over_200k` and `experimental.modes.fast` metadata because the Azure sources confirm the standard pricing and limits, but not those OpenAI-specific fields.

Author: ndeybach

providers/azure-cognitive-services/models/gpt-5.5.toml

@@ -0,0 +1,25 @@
+name = "GPT-5.5"
+family = "gpt"
+release_date = "2026-04-24"
+last_updated = "2026-04-24"
+attachment = true
+reasoning = true
+temperature = false
+knowledge = "2025-12-01"
+tool_call = true
+structured_output = true
+open_weights = false
+
+[cost]
+input = 5.00
+output = 30.00
+cache_read = 0.50
+
+[limit]
+context = 1_050_000
+input = 922_000
+output = 128_000
+
+[modalities]
+input = ["text", "image", "pdf"]
+output = ["text"]

providers/azure/models/gpt-5.5.toml

@@ -0,0 +1,25 @@
+name = "GPT-5.5"
+family = "gpt"
+release_date = "2026-04-24"
+last_updated = "2026-04-24"
+attachment = true
+reasoning = true
+temperature = false
+knowledge = "2025-12-01"
+tool_call = true
+structured_output = true
+open_weights = false
+
+[cost]
+input = 5.00
+output = 30.00
+cache_read = 0.50
+
+[limit]
+context = 1_050_000
+input = 922_000
+output = 128_000
+
+[modalities]
+input = ["text", "image", "pdf"]
+output = ["text"]