fix: update Fireworks model metadata for accounts/fireworks/models/gpt-oss-20b

[github-actions]

fix: update Fireworks model metadata for accounts/fireworks/models/gpt-oss-20b

Closes #556

Source issue: #556

Summary

Field	Value
Provider	fireworks
Primary model	accounts/fireworks/models/gpt-oss-20b
Changed models	accounts/fireworks/models/gpt-oss-20b
Added models	None
Updated models	accounts/fireworks/models/gpt-oss-20b
Verification sources	1 2

Verified metadata

Model	Display name	Parent	Providers	Format	Flavor	Token limits	Pricing	Lifecycle
accounts/fireworks/models/gpt-oss-20b	OpenAI GPT-OSS (20B)		fireworks	openai	chat	input=131072, output=131072	in/out=0.07/0.3 per 1M; cache read=0.04 per 1M	reasoning=true

Verification notes

Verification

• Source 1: Fireworks model page -- confirms $0.07 input / $0.30 output / $0.04 cached input per 1M tokens; context length 131,072 tokens. This is the primary official source for the pricing update.
• Source 2: Fireworks serverless pricing -- general pricing page (model-specific pricing is on individual model pages).

Fields verified from official sources

• input_cost_per_mil_tokens: 0.07 (from Fireworks model page)
• output_cost_per_mil_tokens: 0.3 (from Fireworks model page)
• input_cache_read_cost_per_mil_tokens: 0.04 (from Fireworks model page, listed as "Cached Input")
• max_input_tokens: 131072 (from Fireworks model page, matches existing catalog value)
• max_output_tokens: not explicitly listed on the model page; existing catalog value of 131072 is retained unchanged
• format, flavor, displayName, reasoning, available_providers: unchanged from existing catalog entry, not part of this pricing update

sync_models (LiteLLM) cross-check

The model accounts/fireworks/models/gpt-oss-20b does not exist in the LiteLLM model_prices_and_context_window_backup.json catalog. No cross-check is possible. The official Fireworks model page is the sole authoritative source for these values.

Fields not published or not applicable

• input_cache_write_cost_per_mil_tokens: not listed on the Fireworks model page; omitted
• deprecation_date: not applicable (model is active)
• parent: not applicable (this is a provider-specific deployment, not a dated snapshot of a stable alias)
• multimodal: the model page states "Image Input: Not supported"; existing catalog entry does not set this field, so it is left unchanged

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
accounts/fireworks/models/gpt-oss-20b	input_cost_per_mil_tokens	0.07	0.05	fireworks_ai/accounts/fireworks/models/gpt-oss-20b
accounts/fireworks/models/gpt-oss-20b	output_cost_per_mil_tokens	0.3	0.2	fireworks_ai/accounts/fireworks/models/gpt-oss-20b
accounts/fireworks/models/gpt-oss-20b	input_cache_read_cost_per_mil_tokens	0.04	n/a	fireworks_ai/accounts/fireworks/models/gpt-oss-20b

[vercel]

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

Project	Deployment	Actions	Updated (UTC)
ai-proxy	Ready	Preview, Comment	May 19, 2026 6:45pm