Gap
The model voxtral-small-2507 (Voxtral Small) is listed on Mistral's models page and pricing page as a current, non-deprecated model, but is missing from the proxy catalog. It is a multimodal text+audio chat model that supports the standard /v1/chat/completions endpoint for both text-only and audio-input prompts.
Official sources
Verified fields
- Model ID:
voxtral-small-2507 — confirmed on Mistral models page (listed as voxtral-small-25-07 in overview, voxtral-small-2507 on model card)
- Format:
openai — Mistral convention (all Mistral models in catalog use openai format)
- Flavor:
chat — supports /v1/chat/completions
- Input cost:
$0.10 / 1M tokens (text) — confirmed on pricing page
- Output cost:
$0.40 / 1M tokens — confirmed on pricing page
- Context window (max_input_tokens):
32000 — confirmed on model card page
- Available providers:
["mistral"]
- Multimodal:
true — supports text and audio input
Fields NOT published
- max_output_tokens: Not specified on docs. Should be omitted or left for manual review.
- reasoning: Not documented. Omit.
- Audio input pricing:
$0.004 / minute — listed on pricing page but not a standard catalog field.
Verification Checklist
| Check |
Status |
Detail |
| Cross-source |
✅ |
Model ID confirmed on Mistral models page AND Mistral pricing page |
| Already-fixed |
✅ |
voxtral-small-2507 not present in model_list.json or index.ts |
| ID format |
✅ |
voxtral-small-2507 follows Mistral naming convention — consistent with devstral-small-2507, mistral-small-2603 |
| Duplicate check |
✅ |
No existing open issue covers voxtral-small-2507 |
Local Files Inspected
packages/proxy/schema/model_list.json — confirmed no voxtral entries exist anywhere in the file
packages/proxy/schema/index.ts — confirmed no voxtral entries in AvailableEndpointTypes
{
"kind": "missing_model",
"provider": "mistral",
"models": ["voxtral-small-2507"],
"status": "active",
"model_specs": {
"voxtral-small-2507": {
"format": "openai",
"flavor": "chat",
"multimodal": true,
"input_cost_per_mil_tokens": 0.1,
"output_cost_per_mil_tokens": 0.4,
"max_input_tokens": 32000,
"available_providers": ["mistral"]
}
},
"source_urls": [
"https://docs.mistral.ai/getting-started/models",
"https://mistral.ai/pricing"
]
}
Gap
The model
voxtral-small-2507(Voxtral Small) is listed on Mistral's models page and pricing page as a current, non-deprecated model, but is missing from the proxy catalog. It is a multimodal text+audio chat model that supports the standard/v1/chat/completionsendpoint for both text-only and audio-input prompts.Official sources
Verified fields
voxtral-small-2507— confirmed on Mistral models page (listed asvoxtral-small-25-07in overview,voxtral-small-2507on model card)openai— Mistral convention (all Mistral models in catalog useopenaiformat)chat— supports/v1/chat/completions$0.10 / 1M tokens(text) — confirmed on pricing page$0.40 / 1M tokens— confirmed on pricing page32000— confirmed on model card page["mistral"]true— supports text and audio inputFields NOT published
$0.004 / minute— listed on pricing page but not a standard catalog field.Verification Checklist
voxtral-small-2507not present inmodel_list.jsonorindex.tsvoxtral-small-2507follows Mistral naming convention — consistent withdevstral-small-2507,mistral-small-2603voxtral-small-2507Local Files Inspected
packages/proxy/schema/model_list.json— confirmed novoxtralentries exist anywhere in the filepackages/proxy/schema/index.ts— confirmed novoxtralentries inAvailableEndpointTypes{ "kind": "missing_model", "provider": "mistral", "models": ["voxtral-small-2507"], "status": "active", "model_specs": { "voxtral-small-2507": { "format": "openai", "flavor": "chat", "multimodal": true, "input_cost_per_mil_tokens": 0.1, "output_cost_per_mil_tokens": 0.4, "max_input_tokens": 32000, "available_providers": ["mistral"] } }, "source_urls": [ "https://docs.mistral.ai/getting-started/models", "https://mistral.ai/pricing" ] }