Missing model
The Groq model openai/gpt-oss-safeguard-20b is listed on the official Groq models documentation but has no entry in the proxy model catalog (packages/proxy/schema/model_list.json).
The catalog has the Fireworks variant (accounts/fireworks/models/gpt-oss-safeguard-20b) but not the Groq variant, which uses the openai/ prefix consistent with other Groq-hosted GPT-OSS models already in the catalog:
openai/gpt-oss-120b — present ✓
openai/gpt-oss-20b — present ✓
openai/gpt-oss-safeguard-20b — missing ✗
Model details
| Field |
Value |
Source |
| Model ID |
openai/gpt-oss-safeguard-20b |
Groq models page |
| Status |
Preview |
Groq models page |
| Input price |
$0.075/1M tokens |
Groq models page |
| Output price |
$0.30/1M tokens |
Groq models page |
| Context window |
131,072 tokens |
Groq models page |
| Max completion tokens |
65,536 |
Groq models page |
| Speed |
1,000 tokens/s |
Groq models page |
Suggested entry
"openai/gpt-oss-safeguard-20b": {
"format": "openai",
"flavor": "chat",
"input_cost_per_mil_tokens": 0.075,
"output_cost_per_mil_tokens": 0.3,
"displayName": "GPT-OSS Safeguard (20B)",
"experimental": true,
"max_input_tokens": 131072,
"max_output_tokens": 65536,
"available_providers": [
"groq"
]
}
Verification checklist
Additional note
The Groq docs no longer list meta-llama/llama-4-maverick-17b-128e-instruct, which is currently in the catalog with groq as a provider. This may warrant a separate deprecation review.
Source URLs
{
"kind": "missing_model",
"provider": "groq",
"models": ["openai/gpt-oss-safeguard-20b"],
"status": "preview",
"model_specs": {
"openai/gpt-oss-safeguard-20b": {
"format": "openai",
"flavor": "chat",
"input_cost_per_mil_tokens": 0.075,
"output_cost_per_mil_tokens": 0.3,
"displayName": "GPT-OSS Safeguard (20B)",
"experimental": true,
"max_input_tokens": 131072,
"max_output_tokens": 65536,
"available_providers": ["groq"]
}
},
"source_urls": [
"https://console.groq.com/docs/models"
]
}
Missing model
The Groq model
openai/gpt-oss-safeguard-20bis listed on the official Groq models documentation but has no entry in the proxy model catalog (packages/proxy/schema/model_list.json).The catalog has the Fireworks variant (
accounts/fireworks/models/gpt-oss-safeguard-20b) but not the Groq variant, which uses theopenai/prefix consistent with other Groq-hosted GPT-OSS models already in the catalog:openai/gpt-oss-120b— present ✓openai/gpt-oss-20b— present ✓openai/gpt-oss-safeguard-20b— missing ✗Model details
openai/gpt-oss-safeguard-20bSuggested entry
Verification checklist
accounts/fireworks/models/gpt-oss-safeguard-20b(confirming it is a real, commercially deployed model)openai/gpt-oss-safeguard-20bentry exists inmodel_list.jsonopenai/prefix convention used by other Groq-hosted GPT-OSS models (openai/gpt-oss-120b,openai/gpt-oss-20b)Additional note
The Groq docs no longer list
meta-llama/llama-4-maverick-17b-128e-instruct, which is currently in the catalog withgroqas a provider. This may warrant a separate deprecation review.Source URLs
{ "kind": "missing_model", "provider": "groq", "models": ["openai/gpt-oss-safeguard-20b"], "status": "preview", "model_specs": { "openai/gpt-oss-safeguard-20b": { "format": "openai", "flavor": "chat", "input_cost_per_mil_tokens": 0.075, "output_cost_per_mil_tokens": 0.3, "displayName": "GPT-OSS Safeguard (20B)", "experimental": true, "max_input_tokens": 131072, "max_output_tokens": 65536, "available_providers": ["groq"] } }, "source_urls": [ "https://console.groq.com/docs/models" ] }