github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 00:22:43 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #11432] voxtral #54059

New Issue
Open
opened 2026-04-29 05:09:54 -05:00 by GiteaMirror · 5 comments
GiteaMirror commented 2026-04-29 05:09:54 -05:00
Owner
Copy Link

Originally created by @rlindskog on GitHub (Jul 15, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/11432

Originally created by @rlindskog on GitHub (Jul 15, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/11432
GiteaMirror added the model label 2026-04-29 05:09:54 -05:00
GiteaMirror commented 2026-04-29 05:09:55 -05:00
Author
Owner
Copy Link

@SuperPat45 commented on GitHub (Jul 15, 2025):

This model need to add audio input support to Ollama APIs
For /chat/completions, the Mistral Doc gives this example:

curl --location https://api.mistral.ai/v1/chat/completions \
  --header "Authorization: Bearer $MISTRAL_API_KEY" \
  --header "Content-Type: application/json" \
  --data '{
    "model": "voxtral-mini-2507",
    "messages": [
      {
        "role": "user",
        "content": [
          {
            "type": "input_audio",
            "input_audio": {
              "data": "https://download.samplelib.com/mp3/sample-15s.mp3",
              "format": "mp3"
            }
          },
          {
            "type": "text",
            "text": "What'\''s in this file?"
          }
        ]
      }
    ]
  }'

I suppose it would not be possible to also implement the “standard” /audio/transcriptions API with this model?

<!-- gh-comment-id:3075411655 --> @SuperPat45 commented on GitHub (Jul 15, 2025): This model need to add audio input support to Ollama APIs For /chat/completions, the [Mistral Doc](https://docs.mistral.ai/capabilities/audio/) gives this example: ``` curl --location https://api.mistral.ai/v1/chat/completions \ --header "Authorization: Bearer $MISTRAL_API_KEY" \ --header "Content-Type: application/json" \ --data '{ "model": "voxtral-mini-2507", "messages": [ { "role": "user", "content": [ { "type": "input_audio", "input_audio": { "data": "https://download.samplelib.com/mp3/sample-15s.mp3", "format": "mp3" } }, { "type": "text", "text": "What'\''s in this file?" } ] } ] }' ``` I suppose it would not be possible to also implement the “standard” /audio/transcriptions API with this model?
GiteaMirror commented 2026-04-29 05:09:55 -05:00
Author
Owner
Copy Link

@MyButtermilk commented on GitHub (Jul 21, 2025):

I would rather use Voxtral-Small-latest as it is more accurate and thus competitive with the SOTA. Would it be possible to use it in chat completions mode but only return the prompted transcript via json?

from mistralai import Mistral
client = Mistral(api_key=API_KEY)

messages = [{
"role": "user",
"content": [
{"type": "input_audio",
"input_audio": {"data": audio_url, "format": audio_format}},
{"type": "text",
"text": (
"Transcribe the speech verbatim. "
"No translation. No summary. "
"Return JSON {"text": ""}. "
)},
],
}]

resp = client.chat.complete(
model="voxtral-small-latest",
prompt_mode=None,
temperature=0.0,
response_format={"type": "json_object"},
messages=messages,
max_tokens=4096,
)
transcript = resp.choices[0].message.content # parse JSON -> transcript

<!-- gh-comment-id:3095245155 --> @MyButtermilk commented on GitHub (Jul 21, 2025): I would rather use Voxtral-Small-latest as it is more accurate and thus competitive with the SOTA. Would it be possible to use it in chat completions mode but only return the prompted transcript via json? from mistralai import Mistral client = Mistral(api_key=API_KEY) messages = [{ "role": "user", "content": [ {"type": "input_audio", "input_audio": {"data": audio_url, "format": audio_format}}, {"type": "text", "text": ( "Transcribe the speech verbatim. " "No translation. No summary. " "Return JSON {\"text\": \"<transcript>\"}. " )}, ], }] resp = client.chat.complete( model="voxtral-small-latest", prompt_mode=None, temperature=0.0, response_format={"type": "json_object"}, messages=messages, max_tokens=4096, ) transcript = resp.choices[0].message.content # parse JSON -> transcript
GiteaMirror commented 2026-04-29 05:09:56 -05:00
Author
Owner
Copy Link

@eschmidbauer commented on GitHub (Jul 29, 2025):

voxtral support was recently added to llama.cpp

<!-- gh-comment-id:3132096259 --> @eschmidbauer commented on GitHub (Jul 29, 2025): voxtral support was recently added to [llama.cpp](https://github.com/ggml-org/llama.cpp/commit/00fa15fedc79263fa0285e6a3bbb0cfb3e3878a2)
GiteaMirror commented 2026-04-29 05:09:56 -05:00
Author
Owner
Copy Link

@SEVENID commented on GitHub (Nov 24, 2025):

Is it even planned? >3 months ago llama.cpp received voxtral support.

<!-- gh-comment-id:3572511092 --> @SEVENID commented on GitHub (Nov 24, 2025): Is it even planned? >3 months ago llama.cpp received voxtral support.
GiteaMirror commented 2026-04-29 05:09:57 -05:00
Author
Owner
Copy Link

@SuperPat45 commented on GitHub (Feb 4, 2026):

Mistral just released a new Voxtral mini realtime open weight model
https://huggingface.co/mistralai/Voxtral-Mini-4B-Realtime-2602
https://mistral.ai/news/voxtral-transcribe-2

<!-- gh-comment-id:3849438705 --> @SuperPat45 commented on GitHub (Feb 4, 2026): Mistral just released a new Voxtral mini realtime open weight model https://huggingface.co/mistralai/Voxtral-Mini-4B-Realtime-2602 https://mistral.ai/news/voxtral-transcribe-2
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label model
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#54059
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?