Summary
The accounts/fireworks/models/gpt-oss-20b entry in model_list.json has stale pricing. Fireworks AI's current published pricing differs from the catalog values.
Current values in catalog (line 1548)
| Field |
Current Value |
Correct Value |
input_cost_per_mil_tokens |
0.05 |
0.07 |
output_cost_per_mil_tokens |
0.20 |
0.30 |
input_cache_read_cost_per_mil_tokens |
0.04 |
0.035 |
Verification
- Source 1: Fireworks AI pricing page (https://fireworks.ai/pricing) — lists GPT-OSS 20B at $0.07 input / $0.30 output / $0.035 cached input per 1M tokens
- Source 2: The model is also available via other providers (Groq, Together) at different per-provider prices, confirming it is an active model. The Groq-hosted
openai/gpt-oss-20b in the catalog has $0.075/$0.30 pricing which is closer to Fireworks's current rate than the catalog's Fireworks entry.
- ID format:
accounts/fireworks/models/gpt-oss-20b matches existing Fireworks naming convention
- Recent commits: No recent commits updating this model's pricing
Suggested fix
Update the entry in packages/proxy/schema/model_list.json:
"accounts/fireworks/models/gpt-oss-20b": {
"format": "openai",
"flavor": "chat",
"input_cost_per_mil_tokens": 0.07,
"output_cost_per_mil_tokens": 0.30,
"input_cache_read_cost_per_mil_tokens": 0.035,
"displayName": "OpenAI GPT-OSS (20B)",
"reasoning": true,
"max_input_tokens": 131072,
"max_output_tokens": 131072,
"available_providers": ["fireworks"]
}
Summary
The
accounts/fireworks/models/gpt-oss-20bentry inmodel_list.jsonhas stale pricing. Fireworks AI's current published pricing differs from the catalog values.Current values in catalog (line 1548)
input_cost_per_mil_tokensoutput_cost_per_mil_tokensinput_cache_read_cost_per_mil_tokensVerification
openai/gpt-oss-20bin the catalog has $0.075/$0.30 pricing which is closer to Fireworks's current rate than the catalog's Fireworks entry.accounts/fireworks/models/gpt-oss-20bmatches existing Fireworks naming conventionSuggested fix
Update the entry in
packages/proxy/schema/model_list.json: