fix: add Databricks models databricks-claude-haiku-4-5 +2 more

[github-actions]

fix: add Databricks models databricks-claude-haiku-4-5 +2 more

Closes #640

Source issue: #640

Summary

Field	Value
Provider	databricks
Primary model	databricks-claude-haiku-4-5
Changed models	databricks-claude-haiku-4-5 databricks-claude-opus-4-5 databricks-claude-opus-4-1
Added models	databricks-claude-haiku-4-5 databricks-claude-opus-4-5 databricks-claude-opus-4-1
Updated models	None
Verification sources	1 2

Verified metadata

Model	Display name	Parent	Providers	Format	Flavor	Token limits	Pricing	Lifecycle
databricks-claude-haiku-4-5	Claude 4.5 Haiku		databricks	openai	chat	input=200000, output=not provided	n/a	multimodal=true
databricks-claude-opus-4-5	Claude 4.5 Opus		databricks	openai	chat	input=200000, output=32000	n/a	multimodal=true
databricks-claude-opus-4-1	Claude 4.1 Opus		databricks	openai	chat	input=200000, output=32000	n/a	multimodal=true

Verification notes

Verification

Official sources and fields verified

1. Databricks Supported Models page (https://docs.databricks.com/aws/en/machine-learning/foundation-model-apis/supported-models):

   • Verified model IDs: databricks-claude-haiku-4-5, databricks-claude-opus-4-5, databricks-claude-opus-4-1
   • Verified max_input_tokens: 200000 for Opus 4.5 and Opus 4.1 ("200K token context window")
   • Verified max_output_tokens: 32000 for Opus 4.5 and Opus 4.1 ("32K output token capabilities")
   • Verified multimodal: true for all three (inputs: text, image)
   • Haiku 4.5 context window and max output tokens: NOT PUBLISHED on this page

2. Anthropic model catalog (existing claude-haiku-4-5 entry in model_list.json):

   • Used max_input_tokens: 200000 for databricks-claude-haiku-4-5 — the underlying Anthropic model has a 200K context window, and all other Databricks Claude models (e.g., databricks-claude-3-7-sonnet) use the same context window as the native Anthropic model
   • max_output_tokens omitted for Haiku 4.5 because Databricks does not publish it and may impose a provider-specific cap (e.g., Opus 4.5 is capped at 32K on Databricks vs 64K natively on Anthropic)

3. Existing catalog entry databricks-claude-3-7-sonnet (style reference):

   • Confirmed format: "openai", flavor: "chat" convention for Databricks models
   • Confirmed displayName convention: use model's common name without "Databricks" prefix

sync_models (LiteLLM) cross-check

The LiteLLM model_prices_and_context_window_backup.json contains no entries for any databricks/claude* or databricks-claude* models. Therefore, no numeric cross-check is possible. All proposed values are sourced directly from the Databricks documentation and the underlying Anthropic model specifications.

Fields not published or not applicable

• input_cost_per_mil_tokens / output_cost_per_mil_tokens: Databricks does not publish stable per-model pricing
• max_output_tokens for databricks-claude-haiku-4-5: Not published by Databricks
• reasoning / reasoning_budget: Not verified at API level for Databricks surface
• input_cache_read_cost_per_mil_tokens / input_cache_write_cost_per_mil_tokens: Not applicable (Databricks does not expose Anthropic prompt caching)
• supported_regions: Not applicable (provider is databricks, not vertex)
• parent: Not applicable (these are not dated snapshots or location-scoped variants)
• deprecation_date: Not published

sync_models vs proposed update

sync_models cross-check found differences. Official provider verification was used for the applied values, and sync_models discrepancies are listed below for review.

Model	Field	Proposed update	sync_models	sync_models source models
databricks-claude-haiku-4-5	max_output_tokens	n/a	64000	databricks/databricks-claude-haiku-4-5
databricks-claude-haiku-4-5	input_cost_per_mil_tokens	n/a	1.00002	databricks/databricks-claude-haiku-4-5
databricks-claude-haiku-4-5	output_cost_per_mil_tokens	n/a	5.00003	databricks/databricks-claude-haiku-4-5
databricks-claude-opus-4-5	max_output_tokens	32000	64000	databricks/databricks-claude-opus-4-5
databricks-claude-opus-4-5	input_cost_per_mil_tokens	n/a	5.00003	databricks/databricks-claude-opus-4-5
databricks-claude-opus-4-5	output_cost_per_mil_tokens	n/a	25.00001	databricks/databricks-claude-opus-4-5
databricks-claude-opus-4-1	input_cost_per_mil_tokens	n/a	15.00002	databricks/databricks-claude-opus-4-1
databricks-claude-opus-4-1	output_cost_per_mil_tokens	n/a	75.00003	databricks/databricks-claude-opus-4-1

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
ai-proxy	Ready	Preview, Comment	May 23, 2026 12:30pm