Summary
chat-latest has both max_input_tokens and max_output_tokens wrong in the catalog. OpenAI's documentation shows a 400K input / 128K output context window, but the catalog has 128K / 16K.
| Model ID |
Field |
Catalog |
OpenAI |
chat-latest |
max_input_tokens |
128,000 |
400,000 |
chat-latest |
max_output_tokens |
16,384 |
128,000 |
Pricing ($5/$30 input/output, $0.50 cached input) is correct in the catalog.
Sources
Verification
| Check |
Status |
Detail |
| Cross-source |
✅ |
Model confirmed on model page (context + output limits) and pricing page (model existence + pricing) |
| Already fixed |
✅ |
Confirmed catalog has 128000/16384 at line ~11873 |
| ID format |
✅ |
chat-latest matches existing catalog entry |
| Duplicate check |
✅ |
No open issue covers this gap |
Verification Notes
| Field |
Source |
Notes |
max_input_tokens: 400000 |
chat-latest model page |
"400,000 context window" |
max_output_tokens: 128000 |
Same page |
"128,000 max output tokens" |
| Pricing correct |
Same page + pricing page |
$5/$30/$0.50 cached — matches catalog |
reasoning: true |
Already in catalog |
Correct |
Local Files Inspected
packages/proxy/schema/model_list.json — entry confirmed at line ~11873 with max_input_tokens: 128000 and max_output_tokens: 16384
{
"kind": "stale_metadata",
"provider": "openai",
"models": ["chat-latest"],
"status": "active",
"model_specs": {
"chat-latest": {
"format": "openai",
"flavor": "chat",
"multimodal": true,
"input_cost_per_mil_tokens": 5,
"output_cost_per_mil_tokens": 30,
"input_cache_read_cost_per_mil_tokens": 0.5,
"displayName": "GPT Chat Latest",
"reasoning": true,
"max_input_tokens": 400000,
"max_output_tokens": 128000,
"available_providers": ["openai", "azure"]
}
},
"source_urls": [
"https://developers.openai.com/docs/models/chat-latest",
"https://developers.openai.com/docs/pricing"
]
}
Summary
chat-latesthas bothmax_input_tokensandmax_output_tokenswrong in the catalog. OpenAI's documentation shows a 400K input / 128K output context window, but the catalog has 128K / 16K.chat-latestchat-latestPricing ($5/$30 input/output, $0.50 cached input) is correct in the catalog.
Sources
Verification
chat-latestmatches existing catalog entryVerification Notes
max_input_tokens: 400000max_output_tokens: 128000reasoning: trueLocal Files Inspected
packages/proxy/schema/model_list.json— entry confirmed at line ~11873 withmax_input_tokens: 128000andmax_output_tokens: 16384{ "kind": "stale_metadata", "provider": "openai", "models": ["chat-latest"], "status": "active", "model_specs": { "chat-latest": { "format": "openai", "flavor": "chat", "multimodal": true, "input_cost_per_mil_tokens": 5, "output_cost_per_mil_tokens": 30, "input_cache_read_cost_per_mil_tokens": 0.5, "displayName": "GPT Chat Latest", "reasoning": true, "max_input_tokens": 400000, "max_output_tokens": 128000, "available_providers": ["openai", "azure"] } }, "source_urls": [ "https://developers.openai.com/docs/models/chat-latest", "https://developers.openai.com/docs/pricing" ] }