scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-11-09 08:18:12 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 67

fix(pricing): add the azure gpt-realtime-2 family and price realtime image input per token

Browse source 
Adds azure/gpt-realtime-2, azure/gpt-realtime-2.1 and azure/gpt-realtime-2.1-mini,
which had no cost map entry, reconciles the openai-direct gpt-realtime-2 entry, and
corrects the field the realtime family uses for image input.

Image input on these models is billed per token, and the existing entries carry
that rate as input_cost_per_image. Nothing reads that field for token usage: the
calculator reads input_cost_per_image_token, so image tokens fell through to
input_cost_per_token and billed at the text rate. Measured on a live session that
reported 155 text and 194 image input tokens, this billed $0.00020940 where the
metered rates give $0.00024820. The values were already right; only the field name
was wrong, so this renames it on the three existing azure/gpt-realtime keys and
uses it on the three new ones.

Every rate is taken from Azure's own retail meters rather than from OpenAI's price
card, since Azure prices its OpenAI models independently. For the Global tier, per
1M tokens:

  gpt-realtime-2         text 4.00 in / 24.00 out, image 5.00 in, text cached 0.40
  gpt-realtime-2.1       same, plus audio 32.00 in / 64.00 out, audio cached 0.40
  gpt-realtime-2.1-mini  text 0.60 in / 2.40 out, image 0.80 in, audio 10.00 / 20.00

The Data Zone meters come out at exactly 1.1x Global across every dimension, which
is the expected relationship and a good signal the rows are live rather than stale.
These agree with OpenAI's published realtime rates, but the meters are the
authority for an azure/ key.

The azure entries keep Azure's 32000 input / 4096 output request limits while the
openai-direct entries keep 128000 / 32000, per the review earlier in this PR:
Azure's Realtime API caps a single response at 4096 even though the context window
is 128k, so the two providers legitimately differ here.

Rebuilt on litellm_internal_staging: this branch previously targeted
litellm_oss_daily_2026_07_10, and no daily branch has been cut since 2026_07_20.
Rebuilding also picks up mode: realtime, which staging had already corrected on the
sibling entries. The gpt-realtime-2 reconciliation asked for in review is carried
forward, since staging still has it at 32000/4096 and 16.00 output.
This commit is contained in:
Marty Sullivan 2026-08-16 18:27:57 -04:00
parent 973329e986
commit 06348cca4a
2 changed files with 206 additions and 14 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

110
litellm/model_prices_and_context_window_backup.json
View file

@ -4989,7 +4989,7 @@
"cache_read_input_token_cost": 4e-06,
"deprecation_date": "2027-03-02",
"input_cost_per_audio_token": 3.2e-05,
"input_cost_per_image": 5e-06,
"input_cost_per_image_token": 5e-06,
"input_cost_per_token": 4e-06,
"litellm_provider": "azure",
"max_input_tokens": 32000,
@ -5022,7 +5022,7 @@
"cache_read_input_token_cost": 4e-06,
"deprecation_date": "2027-08-24",
"input_cost_per_audio_token": 3.2e-05,
"input_cost_per_image": 5e-06,
"input_cost_per_image_token": 5e-06,
"input_cost_per_token": 4e-06,
"litellm_provider": "azure",
"max_input_tokens": 32000,
@ -5054,7 +5054,7 @@
"cache_creation_input_audio_token_cost": 3e-07,
"cache_read_input_token_cost": 6e-08,
"input_cost_per_audio_token": 1e-05,
"input_cost_per_image": 8e-07,
"input_cost_per_image_token": 8e-07,
"input_cost_per_token": 6e-07,
"litellm_provider": "azure",
"max_input_tokens": 32000,
@ -26485,12 +26485,12 @@
"input_cost_per_image": 5e-06,
"input_cost_per_token": 4e-06,
"litellm_provider": "openai",
"max_input_tokens": 32000,
"max_output_tokens": 4096,
"max_tokens": 4096,
"max_input_tokens": 128000,
"max_output_tokens": 32000,
"max_tokens": 32000,
"mode": "realtime",
"output_cost_per_audio_token": 6.4e-05,
"output_cost_per_token": 1.6e-05,
"output_cost_per_token": 2.4e-05,
"supported_endpoints": [
"/v1/realtime"
],
@ -48352,5 +48352,101 @@
}
}
]
},
"azure/gpt-realtime-2": {
"cache_creation_input_audio_token_cost": 4e-07,
"cache_read_input_token_cost": 4e-07,
"input_cost_per_audio_token": 3.2e-05,
"input_cost_per_image_token": 5e-06,
"input_cost_per_token": 4e-06,
"litellm_provider": "azure",
"max_input_tokens": 32000,
"max_output_tokens": 4096,
"max_tokens": 4096,
"mode": "realtime",
"output_cost_per_audio_token": 6.4e-05,
"output_cost_per_token": 2.4e-05,
"supported_endpoints": [
"/v1/realtime"
],
"supported_modalities": [
"text",
"image",
"audio"
],
"supported_output_modalities": [
"text",
"audio"
],
"supports_audio_input": true,
"supports_audio_output": true,
"supports_function_calling": true,
"supports_parallel_function_calling": true,
"supports_system_messages": true,
"supports_tool_choice": true
},
"azure/gpt-realtime-2.1": {
"cache_creation_input_audio_token_cost": 4e-07,
"cache_read_input_token_cost": 4e-07,
"input_cost_per_audio_token": 3.2e-05,
"input_cost_per_image_token": 5e-06,
"input_cost_per_token": 4e-06,
"litellm_provider": "azure",
"max_input_tokens": 32000,
"max_output_tokens": 4096,
"max_tokens": 4096,
"mode": "realtime",
"output_cost_per_audio_token": 6.4e-05,
"output_cost_per_token": 2.4e-05,
"supported_endpoints": [
"/v1/realtime"
],
"supported_modalities": [
"text",
"image",
"audio"
],
"supported_output_modalities": [
"text",
"audio"
],
"supports_audio_input": true,
"supports_audio_output": true,
"supports_function_calling": true,
"supports_parallel_function_calling": true,
"supports_system_messages": true,
"supports_tool_choice": true
},
"azure/gpt-realtime-2.1-mini": {
"cache_creation_input_audio_token_cost": 3e-07,
"cache_read_input_token_cost": 6e-08,
"input_cost_per_audio_token": 1e-05,
"input_cost_per_image_token": 8e-07,
"input_cost_per_token": 6e-07,
"litellm_provider": "azure",
"max_input_tokens": 32000,
"max_output_tokens": 4096,
"max_tokens": 4096,
"mode": "realtime",
"output_cost_per_audio_token": 2e-05,
"output_cost_per_token": 2.4e-06,
"supported_endpoints": [
"/v1/realtime"
],
"supported_modalities": [
"text",
"image",
"audio"
],
"supported_output_modalities": [
"text",
"audio"
],
"supports_audio_input": true,
"supports_audio_output": true,
"supports_function_calling": true,
"supports_parallel_function_calling": true,
"supports_system_messages": true,
"supports_tool_choice": true
}
}

110
model_prices_and_context_window.json
View file

@ -4989,7 +4989,7 @@
"cache_read_input_token_cost": 4e-06,
"deprecation_date": "2027-03-02",
"input_cost_per_audio_token": 3.2e-05,
"input_cost_per_image": 5e-06,
"input_cost_per_image_token": 5e-06,
"input_cost_per_token": 4e-06,
"litellm_provider": "azure",
"max_input_tokens": 32000,
@ -5022,7 +5022,7 @@
"cache_read_input_token_cost": 4e-06,
"deprecation_date": "2027-08-24",
"input_cost_per_audio_token": 3.2e-05,
"input_cost_per_image": 5e-06,
"input_cost_per_image_token": 5e-06,
"input_cost_per_token": 4e-06,
"litellm_provider": "azure",
"max_input_tokens": 32000,
@ -5054,7 +5054,7 @@
"cache_creation_input_audio_token_cost": 3e-07,
"cache_read_input_token_cost": 6e-08,
"input_cost_per_audio_token": 1e-05,
"input_cost_per_image": 8e-07,
"input_cost_per_image_token": 8e-07,
"input_cost_per_token": 6e-07,
"litellm_provider": "azure",
"max_input_tokens": 32000,
@ -26485,12 +26485,12 @@
"input_cost_per_image": 5e-06,
"input_cost_per_token": 4e-06,
"litellm_provider": "openai",
"max_input_tokens": 32000,
"max_output_tokens": 4096,
"max_tokens": 4096,
"max_input_tokens": 128000,
"max_output_tokens": 32000,
"max_tokens": 32000,
"mode": "realtime",
"output_cost_per_audio_token": 6.4e-05,
"output_cost_per_token": 1.6e-05,
"output_cost_per_token": 2.4e-05,
"supported_endpoints": [
"/v1/realtime"
],
@ -48352,5 +48352,101 @@
}
}
]
},
"azure/gpt-realtime-2": {
"cache_creation_input_audio_token_cost": 4e-07,
"cache_read_input_token_cost": 4e-07,
"input_cost_per_audio_token": 3.2e-05,
"input_cost_per_image_token": 5e-06,
"input_cost_per_token": 4e-06,
"litellm_provider": "azure",
"max_input_tokens": 32000,
"max_output_tokens": 4096,
"max_tokens": 4096,
"mode": "realtime",
"output_cost_per_audio_token": 6.4e-05,
"output_cost_per_token": 2.4e-05,
"supported_endpoints": [
"/v1/realtime"
],
"supported_modalities": [
"text",
"image",
"audio"
],
"supported_output_modalities": [
"text",
"audio"
],
"supports_audio_input": true,
"supports_audio_output": true,
"supports_function_calling": true,
"supports_parallel_function_calling": true,
"supports_system_messages": true,
"supports_tool_choice": true
},
"azure/gpt-realtime-2.1": {
"cache_creation_input_audio_token_cost": 4e-07,
"cache_read_input_token_cost": 4e-07,
"input_cost_per_audio_token": 3.2e-05,
"input_cost_per_image_token": 5e-06,
"input_cost_per_token": 4e-06,
"litellm_provider": "azure",
"max_input_tokens": 32000,
"max_output_tokens": 4096,
"max_tokens": 4096,
"mode": "realtime",
"output_cost_per_audio_token": 6.4e-05,
"output_cost_per_token": 2.4e-05,
"supported_endpoints": [
"/v1/realtime"
],
"supported_modalities": [
"text",
"image",
"audio"
],
"supported_output_modalities": [
"text",
"audio"
],
"supports_audio_input": true,
"supports_audio_output": true,
"supports_function_calling": true,
"supports_parallel_function_calling": true,
"supports_system_messages": true,
"supports_tool_choice": true
},
"azure/gpt-realtime-2.1-mini": {
"cache_creation_input_audio_token_cost": 3e-07,
"cache_read_input_token_cost": 6e-08,
"input_cost_per_audio_token": 1e-05,
"input_cost_per_image_token": 8e-07,
"input_cost_per_token": 6e-07,
"litellm_provider": "azure",
"max_input_tokens": 32000,
"max_output_tokens": 4096,
"max_tokens": 4096,
"mode": "realtime",
"output_cost_per_audio_token": 2e-05,
"output_cost_per_token": 2.4e-06,
"supported_endpoints": [
"/v1/realtime"
],
"supported_modalities": [
"text",
"image",
"audio"
],
"supported_output_modalities": [
"text",
"audio"
],
"supports_audio_input": true,
"supports_audio_output": true,
"supports_function_calling": true,
"supports_parallel_function_calling": true,
"supports_system_messages": true,
"supports_tool_choice": true
}
}