branch: elpa/gptel
commit 553a4035004ef9153df8fcc3fda5593351f55fb7
Author: Karthik Chikmagalur <karthikchikmaga...@gmail.com>
Commit: Karthik Chikmagalur <karthikchikmaga...@gmail.com>
gptel-openai: Add support for o3-mini
* gptel-openai.el (gptel--request-data): Support o3-mini
* gptel.el (gptel--openai-models): Add details for o3-mini
---
gptel-openai.el | 2 +-
gptel.el | 14 +++++++++++---
2 files changed, 12 insertions(+), 4 deletions(-)
diff --git a/gptel-openai.el b/gptel-openai.el
index fd67ac8d0a..d5746f6ec8 100644
--- a/gptel-openai.el
+++ b/gptel-openai.el
@@ -283,7 +283,7 @@ Mutate state INFO with response metadata."
(when gptel-max-tokens
;; HACK: The OpenAI API has deprecated max_tokens, but we still need it
;; for OpenAI-compatible APIs like GPT4All (#485)
- (plist-put prompts-plist (if (memq gptel-model '(o1 o1-preview o1-mini))
+ (plist-put prompts-plist (if (memq gptel-model '(o1 o1-preview o1-mini
o3-mini))
:max_completion_tokens :max_tokens)
gptel-max-tokens))
;; Merge request params with model and backend params.
diff --git a/gptel.el b/gptel.el
index dd726fcde6..d6c8801ce6 100644
--- a/gptel.el
+++ b/gptel.el
@@ -555,7 +555,7 @@ To set the temperature for a chat session interactively call
:cutoff-date "2023-12")
(o1
:description "Reasoning model designed to solve hard problems across
domains"
- :capabilities (nosystem media)
+ :capabilities (nosystem media reasoning)
:mime-types ("image/jpeg" "image/png" "image/gif" "image/webp")
:context-window 200
:input-cost 15
@@ -570,7 +570,7 @@ To set the temperature for a chat session interactively call
:input-cost 15
:output-cost 60
:cutoff-date "2023-10"
- :capabilities (nosystem)
+ :capabilities (nosystem reasoning)
:request-params (:stream :json-false))
(o1-mini
:description "Faster and cheaper reasoning model good at coding, math,
and science"
@@ -578,7 +578,15 @@ To set the temperature for a chat session interactively
call
:input-cost 3
:output-cost 12
:cutoff-date "2023-10"
- :capabilities (nosystem)
+ :capabilities (nosystem reasoning)
+ :request-params (:stream :json-false))
+ (o3-mini
+ :description "High intelligence at the same cost and latency targets of
o1-mini"
+ :context-window 200
+ :input-cost 3
+ :output-cost 12
+ :cutoff-date "2023-10"
+ :capabilities (nosystem reasoning)
:request-params (:stream :json-false))
;; limited information available
(gpt-4-32k
