1version: 1
2refreshedAt: '2026-08-15T17:14:10.075Z'
3models:
4 allenai/olmOCR-2-7B-1025:
5 default:
6 max_tokens:
7 verdict: supported
8 source: prose
9 lastChecked: '2026-06-11'
10 evidencePath: /Users/eshao/mnt/mdr/skills/models-litellm/assets/config.yml localRoutes DMR OpenAI-compatible route
11 type: integer
12 note: min 16; An upper bound for the number of tokens that can be generated for a response, including visible output tokens and [reasoning tokens](/docs/guides/reasoning).; alias of max_output_tokens
13 max_completion_tokens:
14 verdict: supported
15 source: prose
16 lastChecked: '2026-06-11'
17 evidencePath: /Users/eshao/mnt/mdr/skills/models-litellm/assets/models.yml parameterMapping max_completion_tokens -> max_tokens
18 type: integer
19 note: min 16; An upper bound for the number of tokens that can be generated for a response, including visible output tokens and [reasoning tokens](/docs/guides/reasoning).; alias of max_output_tokens
20 metadata:
21 verdict: supported
22 source: schema
23 lastChecked: '2026-06-11'
24 evidencePath: https://raw.githubusercontent.com/openai/openai-openapi/master/openapi.yaml
25 note: Set of 16 key-value pairs that can be attached to an object. This can be
26 logprobs:
27 verdict: supported
28 source: eval
29 lastChecked: '2026-07-15'
30 evidencePath: /Users/eshao/mnt/data/models-catalog/conformance/allenai_olmOCR-2-7B-1025/default/logprobs.json
31 type: boolean
32 top_logprobs:
33 verdict: supported
34 source: eval
35 lastChecked: '2026-07-15'
36 evidencePath: /Users/eshao/mnt/data/models-catalog/conformance/allenai_olmOCR-2-7B-1025/default/top_logprobs.json
37 type: integer
38 note: min 0; max 20; An integer between 0 and 20 specifying the maximum number of most likely
39 temperature:
40 verdict: supported
41 source: schema
42 lastChecked: '2026-06-11'
43 evidencePath: https://raw.githubusercontent.com/openai/openai-openapi/master/openapi.yaml
44 note: min 0; max 2; default 1; What sampling temperature to use, between 0 and 2. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic.
45 top_p:
46 verdict: supported
47 source: schema
48 lastChecked: '2026-06-11'
49 evidencePath: https://raw.githubusercontent.com/openai/openai-openapi/master/openapi.yaml
50 note: min 0; max 1; default 1; An alternative to sampling with temperature, called nucleus sampling,
51 user:
52 verdict: supported
53 source: schema
54 lastChecked: '2026-06-11'
55 evidencePath: https://raw.githubusercontent.com/openai/openai-openapi/master/openapi.yaml
56 type: string
57 note: This field is being replaced by `safety_identifier` and `prompt_cache_key`. Use `prompt_cache_key` instead to maintain caching optimizations.
58 safety_identifier:
59 verdict: supported
60 source: schema
61 lastChecked: '2026-06-11'
62 evidencePath: https://raw.githubusercontent.com/openai/openai-openapi/master/openapi.yaml
63 type: string
64 note: A stable identifier used to help detect users of your application that may be violating OpenAI's usage policies.
65 prompt_cache_key:
66 verdict: supported
67 source: schema
68 lastChecked: '2026-06-11'
69 evidencePath: https://raw.githubusercontent.com/openai/openai-openapi/master/openapi.yaml
70 type: string
71 note: Used by OpenAI to cache responses for similar requests to optimize your cache hit rates. Replaces the `user` field. [Learn more](/docs/guides/prompt-caching).
72 service_tier:
73 verdict: supported
74 source: schema
75 lastChecked: '2026-06-11'
76 evidencePath: https://raw.githubusercontent.com/openai/openai-openapi/master/openapi.yaml
77 type: enum
78 values:
79 - auto
80 - default
81 - flex
82 - scale
83 - priority
84 - fast
85 - ultrafast
86 valuesComplete: true
87 note: default auto; Specifies the processing type used for serving the request.
88 prompt_cache_retention:
89 verdict: supported
90 source: schema
91 lastChecked: '2026-06-11'
92 evidencePath: https://raw.githubusercontent.com/openai/openai-openapi/master/openapi.yaml
93 type: enum
94 values:
95 - in_memory
96 - 24h
97 valuesComplete: true
98 note: The retention policy for the prompt cache. Set to `24h` to enable extended prompt caching, which keeps cached prefixes active for longer, up to a maximum of 24 hours. [Learn more](/docs/guides/prompt-caching#prompt-cache-retention).
99 previous_response_id:
100 verdict: supported
101 source: schema
102 lastChecked: '2026-06-11'
103 evidencePath: https://raw.githubusercontent.com/openai/openai-openapi/master/openapi.yaml
104 type: string
105 note: The unique ID of the previous response to the model. Use this to
106 model:
107 verdict: supported
108 source: schema
109 lastChecked: '2026-06-11'
110 evidencePath: https://raw.githubusercontent.com/openai/openai-openapi/master/openapi.yaml
111 type: string
112 reasoning:
113 verdict: supported
114 source: schema
115 lastChecked: '2026-06-11'
116 evidencePath: https://raw.githubusercontent.com/openai/openai-openapi/master/openapi.yaml
117 note: '**gpt-5 and o-series models only**'
118 background:
119 verdict: supported
120 source: schema
121 lastChecked: '2026-06-11'
122 evidencePath: https://raw.githubusercontent.com/openai/openai-openapi/master/openapi.yaml
123 type: boolean
124 note: default false; Whether to run the model response in the background.
125 max_tool_calls:
126 verdict: supported
127 source: schema
128 lastChecked: '2026-06-11'
129 evidencePath: https://raw.githubusercontent.com/openai/openai-openapi/master/openapi.yaml
130 type: integer
131 note: The maximum number of total calls to built-in tools that can be processed in a response. This maximum number applies across all built-in tool calls, not per individual tool. Any further attempts to call a tool by the model will be ignored.
132 text:
133 verdict: supported
134 source: schema
135 lastChecked: '2026-06-11'
136 evidencePath: https://raw.githubusercontent.com/openai/openai-openapi/master/openapi.yaml
137 note: Configuration options for a text response from the model. Can be plain
138 tools:
139 verdict: supported
140 source: schema
141 lastChecked: '2026-06-11'
142 evidencePath: https://raw.githubusercontent.com/openai/openai-openapi/master/openapi.yaml
143 note: An array of tools the model may call while generating a response. You
144 tool_choice:
145 verdict: supported
146 source: schema
147 lastChecked: '2026-06-11'
148 evidencePath: https://raw.githubusercontent.com/openai/openai-openapi/master/openapi.yaml
149 type: enum
150 values:
151 - none
152 - auto
153 - required
154 valuesComplete: true
155 note: Controls which (if any) tool is called by the model.
156 prompt:
157 verdict: supported
158 source: schema
159 lastChecked: '2026-06-11'
160 evidencePath: https://raw.githubusercontent.com/openai/openai-openapi/master/openapi.yaml
161 note: Reference to a prompt template and its variables.
162 truncation:
163 verdict: supported
164 source: schema
165 lastChecked: '2026-06-11'
166 evidencePath: https://raw.githubusercontent.com/openai/openai-openapi/master/openapi.yaml
167 type: enum
168 values:
169 - auto
170 - disabled
171 valuesComplete: true
172 note: default disabled; The truncation strategy to use for the model response.
173 input:
174 verdict: supported
175