raif 1.3.0 → 1.5.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (206) hide show
  1. checksums.yaml +4 -4
  2. data/README.md +8 -7
  3. data/app/assets/builds/raif.css +4 -1
  4. data/app/assets/builds/raif_admin.css +52 -2
  5. data/app/assets/builds/raif_admin_sprockets.js +2709 -0
  6. data/app/assets/javascript/raif/admin/copy_to_clipboard_controller.js +132 -0
  7. data/app/assets/javascript/raif/admin/cost_estimate_controller.js +80 -0
  8. data/app/assets/javascript/raif/admin/judge_config_controller.js +23 -0
  9. data/app/assets/javascript/raif/admin/select_all_checkboxes_controller.js +33 -0
  10. data/app/assets/javascript/raif/admin/sortable_table_controller.js +51 -0
  11. data/app/assets/javascript/raif/admin/table_search_controller.js +15 -0
  12. data/app/assets/javascript/raif/admin/tom_select_controller.js +33 -0
  13. data/app/assets/javascript/raif/controllers/conversations_controller.js +1 -1
  14. data/app/assets/javascript/raif_admin.js +23 -0
  15. data/app/assets/javascript/raif_admin_sprockets.js +24 -0
  16. data/app/assets/stylesheets/raif/admin/conversation.scss +16 -0
  17. data/app/assets/stylesheets/raif/conversations.scss +3 -0
  18. data/app/assets/stylesheets/raif.scss +2 -1
  19. data/app/assets/stylesheets/raif_admin.scss +50 -1
  20. data/app/controllers/raif/admin/agents_controller.rb +27 -1
  21. data/app/controllers/raif/admin/application_controller.rb +16 -0
  22. data/app/controllers/raif/admin/configs_controller.rb +95 -0
  23. data/app/controllers/raif/admin/llms_controller.rb +27 -0
  24. data/app/controllers/raif/admin/model_completions_controller.rb +24 -1
  25. data/app/controllers/raif/admin/model_tool_invocations_controller.rb +7 -1
  26. data/app/controllers/raif/admin/prompt_studio/agents_controller.rb +25 -0
  27. data/app/controllers/raif/admin/prompt_studio/base_controller.rb +32 -0
  28. data/app/controllers/raif/admin/prompt_studio/batch_runs_controller.rb +102 -0
  29. data/app/controllers/raif/admin/prompt_studio/conversations_controller.rb +25 -0
  30. data/app/controllers/raif/admin/prompt_studio/tasks_controller.rb +64 -0
  31. data/app/controllers/raif/admin/stats/model_tool_invocations_controller.rb +21 -0
  32. data/app/controllers/raif/admin/stats/tasks_controller.rb +15 -6
  33. data/app/controllers/raif/admin/stats_controller.rb +32 -3
  34. data/app/controllers/raif/admin/tasks_controller.rb +5 -0
  35. data/app/controllers/raif/conversation_entries_controller.rb +1 -0
  36. data/app/controllers/raif/conversations_controller.rb +10 -2
  37. data/app/helpers/raif/application_helper.rb +40 -0
  38. data/app/jobs/raif/conversation_entry_job.rb +8 -6
  39. data/app/jobs/raif/prompt_studio_batch_run_item_job.rb +11 -0
  40. data/app/jobs/raif/prompt_studio_batch_run_job.rb +15 -0
  41. data/app/jobs/raif/prompt_studio_task_run_job.rb +36 -0
  42. data/app/models/raif/admin/task_stat.rb +7 -0
  43. data/app/models/raif/agent.rb +98 -6
  44. data/app/models/raif/agents/native_tool_calling_agent.rb +179 -52
  45. data/app/models/raif/application_record.rb +18 -0
  46. data/app/models/raif/concerns/agent_inference_stats.rb +35 -0
  47. data/app/models/raif/concerns/has_prompt_templates.rb +88 -0
  48. data/app/models/raif/concerns/has_runtime_duration.rb +41 -0
  49. data/app/models/raif/concerns/json_schema_definition.rb +54 -6
  50. data/app/models/raif/concerns/llm_prompt_caching.rb +20 -0
  51. data/app/models/raif/concerns/llms/anthropic/message_formatting.rb +34 -0
  52. data/app/models/raif/concerns/llms/anthropic/response_tool_calls.rb +24 -0
  53. data/app/models/raif/concerns/llms/anthropic/tool_formatting.rb +8 -0
  54. data/app/models/raif/concerns/llms/bedrock/message_formatting.rb +43 -0
  55. data/app/models/raif/concerns/llms/bedrock/response_tool_calls.rb +26 -0
  56. data/app/models/raif/concerns/llms/bedrock/tool_formatting.rb +8 -0
  57. data/app/models/raif/concerns/llms/google/message_formatting.rb +112 -0
  58. data/app/models/raif/concerns/llms/google/response_tool_calls.rb +32 -0
  59. data/app/models/raif/concerns/llms/google/tool_formatting.rb +76 -0
  60. data/app/models/raif/concerns/llms/message_formatting.rb +41 -5
  61. data/app/models/raif/concerns/llms/open_ai/json_schema_validation.rb +3 -3
  62. data/app/models/raif/concerns/llms/open_ai_completions/message_formatting.rb +22 -0
  63. data/app/models/raif/concerns/llms/open_ai_completions/response_tool_calls.rb +22 -0
  64. data/app/models/raif/concerns/llms/open_ai_completions/tool_formatting.rb +8 -0
  65. data/app/models/raif/concerns/llms/open_ai_responses/message_formatting.rb +17 -0
  66. data/app/models/raif/concerns/llms/open_ai_responses/response_tool_calls.rb +26 -0
  67. data/app/models/raif/concerns/llms/open_ai_responses/tool_formatting.rb +8 -0
  68. data/app/models/raif/concerns/provider_managed_tool_calls.rb +162 -0
  69. data/app/models/raif/concerns/run_with.rb +127 -0
  70. data/app/models/raif/conversation.rb +112 -8
  71. data/app/models/raif/conversation_entry.rb +38 -4
  72. data/app/models/raif/embedding_model.rb +2 -1
  73. data/app/models/raif/embedding_models/bedrock.rb +10 -1
  74. data/app/models/raif/embedding_models/google.rb +37 -0
  75. data/app/models/raif/embedding_models/open_ai.rb +1 -1
  76. data/app/models/raif/evals/llm_judge.rb +70 -0
  77. data/{lib → app/models}/raif/evals/llm_judges/binary.rb +41 -3
  78. data/{lib → app/models}/raif/evals/llm_judges/comparative.rb +41 -3
  79. data/{lib → app/models}/raif/evals/llm_judges/scored.rb +39 -1
  80. data/{lib → app/models}/raif/evals/llm_judges/summarization.rb +40 -2
  81. data/app/models/raif/llm.rb +104 -4
  82. data/app/models/raif/llms/anthropic.rb +32 -22
  83. data/app/models/raif/llms/bedrock.rb +64 -24
  84. data/app/models/raif/llms/google.rb +166 -0
  85. data/app/models/raif/llms/open_ai_base.rb +23 -5
  86. data/app/models/raif/llms/open_ai_completions.rb +14 -12
  87. data/app/models/raif/llms/open_ai_responses.rb +14 -17
  88. data/app/models/raif/llms/open_router.rb +16 -15
  89. data/app/models/raif/model_completion.rb +103 -1
  90. data/app/models/raif/model_tool.rb +55 -5
  91. data/app/models/raif/model_tool_invocation.rb +68 -6
  92. data/app/models/raif/model_tools/agent_final_answer.rb +2 -7
  93. data/app/models/raif/model_tools/provider_managed/code_execution.rb +4 -0
  94. data/app/models/raif/model_tools/provider_managed/image_generation.rb +4 -0
  95. data/app/models/raif/model_tools/provider_managed/web_search.rb +4 -0
  96. data/app/models/raif/prompt_studio_batch_run.rb +155 -0
  97. data/app/models/raif/prompt_studio_batch_run_item.rb +220 -0
  98. data/app/models/raif/streaming_responses/bedrock.rb +60 -1
  99. data/app/models/raif/streaming_responses/google.rb +71 -0
  100. data/app/models/raif/task.rb +85 -18
  101. data/app/models/raif/user_tool_invocation.rb +19 -0
  102. data/app/views/layouts/raif/admin.html.erb +43 -2
  103. data/app/views/raif/admin/agents/_agent.html.erb +9 -0
  104. data/app/views/raif/admin/agents/_conversation_message.html.erb +28 -6
  105. data/app/views/raif/admin/agents/index.html.erb +50 -0
  106. data/app/views/raif/admin/agents/show.html.erb +50 -1
  107. data/app/views/raif/admin/configs/show.html.erb +117 -0
  108. data/app/views/raif/admin/conversations/_conversation_entry.html.erb +29 -34
  109. data/app/views/raif/admin/conversations/show.html.erb +2 -0
  110. data/app/views/raif/admin/llms/index.html.erb +110 -0
  111. data/app/views/raif/admin/model_completions/_model_completion.html.erb +10 -5
  112. data/app/views/raif/admin/model_completions/index.html.erb +40 -1
  113. data/app/views/raif/admin/model_completions/show.html.erb +256 -84
  114. data/app/views/raif/admin/model_tool_invocations/index.html.erb +22 -1
  115. data/app/views/raif/admin/model_tool_invocations/show.html.erb +18 -0
  116. data/app/views/raif/admin/model_tools/_list.html.erb +16 -0
  117. data/app/views/raif/admin/model_tools/_model_tool.html.erb +36 -0
  118. data/app/views/raif/admin/prompt_studio/agents/index.html.erb +56 -0
  119. data/app/views/raif/admin/prompt_studio/agents/show.html.erb +57 -0
  120. data/app/views/raif/admin/prompt_studio/batch_runs/_batch_run_item.html.erb +54 -0
  121. data/app/views/raif/admin/prompt_studio/batch_runs/_judge_config_fields.html.erb +76 -0
  122. data/app/views/raif/admin/prompt_studio/batch_runs/_judge_detail_modal.html.erb +27 -0
  123. data/app/views/raif/admin/prompt_studio/batch_runs/_modal.html.erb +35 -0
  124. data/app/views/raif/admin/prompt_studio/batch_runs/_progress.html.erb +78 -0
  125. data/app/views/raif/admin/prompt_studio/batch_runs/show.html.erb +49 -0
  126. data/app/views/raif/admin/prompt_studio/conversations/index.html.erb +48 -0
  127. data/app/views/raif/admin/prompt_studio/conversations/show.html.erb +36 -0
  128. data/app/views/raif/admin/prompt_studio/shared/_nav_tabs.html.erb +17 -0
  129. data/app/views/raif/admin/prompt_studio/shared/_prompt_comparison.html.erb +87 -0
  130. data/app/views/raif/admin/prompt_studio/shared/_type_filter.html.erb +54 -0
  131. data/app/views/raif/admin/prompt_studio/tasks/_task_result.html.erb +145 -0
  132. data/app/views/raif/admin/prompt_studio/tasks/_task_row.html.erb +12 -0
  133. data/app/views/raif/admin/prompt_studio/tasks/_task_type_filter.html.erb +58 -0
  134. data/app/views/raif/admin/prompt_studio/tasks/_tasks_table.html.erb +22 -0
  135. data/app/views/raif/admin/prompt_studio/tasks/index.html.erb +35 -0
  136. data/app/views/raif/admin/prompt_studio/tasks/show.html.erb +19 -0
  137. data/app/views/raif/admin/stats/_stats_tile.html.erb +34 -0
  138. data/app/views/raif/admin/stats/index.html.erb +71 -88
  139. data/app/views/raif/admin/stats/model_tool_invocations/index.html.erb +43 -0
  140. data/app/views/raif/admin/stats/tasks/index.html.erb +20 -6
  141. data/app/views/raif/admin/tasks/_task.html.erb +1 -0
  142. data/app/views/raif/admin/tasks/index.html.erb +23 -6
  143. data/app/views/raif/admin/tasks/show.html.erb +56 -3
  144. data/app/views/raif/conversation_entries/_form.html.erb +3 -0
  145. data/app/views/raif/conversation_entries/_message.html.erb +10 -6
  146. data/app/views/raif/conversations/_conversation.html.erb +10 -0
  147. data/app/views/raif/conversations/_entry_processed.turbo_stream.erb +12 -0
  148. data/app/views/raif/conversations/index.html.erb +23 -0
  149. data/config/importmap.rb +8 -0
  150. data/config/locales/admin.en.yml +161 -1
  151. data/config/locales/en.yml +67 -4
  152. data/config/routes.rb +10 -0
  153. data/db/migrate/20250904194456_add_generating_entry_response_to_raif_conversations.rb +7 -0
  154. data/db/migrate/20250911125234_add_source_to_raif_tasks.rb +7 -0
  155. data/db/migrate/20251020005853_add_source_to_raif_agents.rb +7 -0
  156. data/db/migrate/20251020011346_rename_task_run_args_to_run_with.rb +7 -0
  157. data/db/migrate/20251020011405_add_run_with_to_raif_agents.rb +13 -0
  158. data/db/migrate/20251024160119_add_llm_messages_max_length_to_raif_conversations.rb +14 -0
  159. data/db/migrate/20251124185033_add_provider_tool_call_id_to_raif_model_tool_invocations.rb +7 -0
  160. data/db/migrate/20251128202941_add_tool_choice_to_raif_model_completions.rb +7 -0
  161. data/db/migrate/20260118144846_add_source_to_raif_conversations.rb +7 -0
  162. data/db/migrate/20260119000000_add_failure_tracking_to_raif_model_completions.rb +10 -0
  163. data/db/migrate/20260119000001_add_completed_at_to_raif_model_completions.rb +8 -0
  164. data/db/migrate/20260119000002_add_started_at_to_raif_model_completions.rb +8 -0
  165. data/db/migrate/20260307000000_add_prompt_studio_run_to_raif_tasks.rb +7 -0
  166. data/db/migrate/20260308000000_create_raif_prompt_studio_batch_runs.rb +27 -0
  167. data/db/migrate/20260308000001_create_raif_prompt_studio_batch_run_items.rb +24 -0
  168. data/db/migrate/20260407000000_add_cache_token_columns_to_raif_model_completions.rb +8 -0
  169. data/lib/generators/raif/agent/agent_generator.rb +18 -0
  170. data/lib/generators/raif/agent/templates/agent.rb.tt +7 -5
  171. data/lib/generators/raif/agent/templates/application_agent.rb.tt +1 -1
  172. data/lib/generators/raif/agent/templates/system_prompt.erb.tt +3 -0
  173. data/lib/generators/raif/conversation/conversation_generator.rb +19 -1
  174. data/lib/generators/raif/conversation/templates/conversation.rb.tt +6 -0
  175. data/lib/generators/raif/conversation/templates/system_prompt.erb.tt +4 -0
  176. data/lib/generators/raif/install/templates/initializer.rb +117 -8
  177. data/lib/generators/raif/task/task_generator.rb +18 -0
  178. data/lib/generators/raif/task/templates/prompt.erb.tt +4 -0
  179. data/lib/generators/raif/task/templates/task.rb.tt +10 -9
  180. data/lib/raif/configuration.rb +47 -2
  181. data/lib/raif/embedding_model_registry.rb +8 -0
  182. data/lib/raif/engine.rb +24 -1
  183. data/lib/raif/errors/blank_response_error.rb +8 -0
  184. data/lib/raif/errors/instance_dependent_schema_error.rb +8 -0
  185. data/lib/raif/errors/prompt_template_error.rb +15 -0
  186. data/lib/raif/errors/streaming_error.rb +6 -3
  187. data/lib/raif/errors.rb +3 -0
  188. data/lib/raif/evals/run.rb +1 -0
  189. data/lib/raif/evals.rb +0 -6
  190. data/lib/raif/json_schema_builder.rb +14 -0
  191. data/lib/raif/llm_registry.rb +433 -42
  192. data/lib/raif/messages.rb +180 -0
  193. data/lib/raif/prompt_studio_comparison_builder.rb +138 -0
  194. data/lib/raif/token_estimator.rb +28 -0
  195. data/lib/raif/version.rb +1 -1
  196. data/lib/raif.rb +11 -0
  197. data/lib/tasks/annotate_rb.rake +10 -0
  198. data/spec/support/rspec_helpers.rb +15 -9
  199. data/spec/support/test_task.rb +9 -0
  200. data/spec/support/test_template_task.rb +41 -0
  201. metadata +108 -15
  202. data/app/models/raif/agents/re_act_agent.rb +0 -127
  203. data/app/models/raif/agents/re_act_step.rb +0 -32
  204. data/app/models/raif/concerns/task_run_args.rb +0 -62
  205. data/lib/raif/evals/llm_judge.rb +0 -32
  206. /data/{lib → app/models}/raif/evals/scoring_rubric.rb +0 -0
@@ -41,6 +41,69 @@ module Raif
41
41
 
42
42
  def self.default_llms
43
43
  open_ai_models = [
44
+ {
45
+ key: :open_ai_gpt_5_4,
46
+ api_name: "gpt-5.4",
47
+ input_token_cost: 2.5 / 1_000_000,
48
+ output_token_cost: 15.0 / 1_000_000,
49
+ model_provider_settings: { supports_temperature: false },
50
+ },
51
+ {
52
+ key: :open_ai_gpt_5_2,
53
+ api_name: "gpt-5.2",
54
+ input_token_cost: 1.75 / 1_000_000,
55
+ output_token_cost: 14.0 / 1_000_000,
56
+ model_provider_settings: { supports_temperature: false },
57
+ },
58
+ {
59
+ key: :open_ai_gpt_5_3,
60
+ api_name: "gpt-5.3",
61
+ input_token_cost: 1.75 / 1_000_000,
62
+ output_token_cost: 14.0 / 1_000_000,
63
+ model_provider_settings: { supports_temperature: false },
64
+ },
65
+ {
66
+ key: :open_ai_gpt_5_1,
67
+ api_name: "gpt-5.1",
68
+ input_token_cost: 1.25 / 1_000_000,
69
+ output_token_cost: 10.0 / 1_000_000,
70
+ model_provider_settings: { supports_temperature: false },
71
+ },
72
+ {
73
+ key: :open_ai_gpt_5,
74
+ api_name: "gpt-5",
75
+ input_token_cost: 1.25 / 1_000_000,
76
+ output_token_cost: 10.0 / 1_000_000,
77
+ model_provider_settings: { supports_temperature: false },
78
+ },
79
+ {
80
+ key: :open_ai_gpt_5_4_mini,
81
+ api_name: "gpt-5.4-mini",
82
+ input_token_cost: 0.75 / 1_000_000,
83
+ output_token_cost: 4.5 / 1_000_000,
84
+ model_provider_settings: { supports_temperature: false },
85
+ },
86
+ {
87
+ key: :open_ai_gpt_5_4_nano,
88
+ api_name: "gpt-5.4-nano",
89
+ input_token_cost: 0.20 / 1_000_000,
90
+ output_token_cost: 1.25 / 1_000_000,
91
+ model_provider_settings: { supports_temperature: false },
92
+ },
93
+ {
94
+ key: :open_ai_gpt_5_mini,
95
+ api_name: "gpt-5-mini",
96
+ input_token_cost: 0.25 / 1_000_000,
97
+ output_token_cost: 2.0 / 1_000_000,
98
+ model_provider_settings: { supports_temperature: false },
99
+ },
100
+ {
101
+ key: :open_ai_gpt_5_nano,
102
+ api_name: "gpt-5-nano",
103
+ input_token_cost: 0.05 / 1_000_000,
104
+ output_token_cost: 0.4 / 1_000_000,
105
+ model_provider_settings: { supports_temperature: false },
106
+ },
44
107
  {
45
108
  key: :open_ai_gpt_4o_mini,
46
109
  api_name: "gpt-4o-mini",
@@ -88,8 +151,8 @@ module Raif
88
151
  {
89
152
  key: :open_ai_o1_mini,
90
153
  api_name: "o1-mini",
91
- input_token_cost: 1.5 / 1_000_000,
92
- output_token_cost: 6.0 / 1_000_000,
154
+ input_token_cost: 1.1 / 1_000_000,
155
+ output_token_cost: 4.4 / 1_000_000,
93
156
  model_provider_settings: { supports_temperature: false },
94
157
  },
95
158
  {
@@ -113,27 +176,6 @@ module Raif
113
176
  output_token_cost: 4.4 / 1_000_000,
114
177
  model_provider_settings: { supports_temperature: false },
115
178
  },
116
- {
117
- key: :open_ai_gpt_5,
118
- api_name: "gpt-5",
119
- input_token_cost: 1.25 / 1_000_000,
120
- output_token_cost: 10.0 / 1_000_000,
121
- model_provider_settings: { supports_temperature: false },
122
- },
123
- {
124
- key: :open_ai_gpt_5_mini,
125
- api_name: "gpt-5-mini",
126
- input_token_cost: 0.25 / 1_000_000,
127
- output_token_cost: 2.0 / 1_000_000,
128
- model_provider_settings: { supports_temperature: false },
129
- },
130
- {
131
- key: :open_ai_gpt_5_nano,
132
- api_name: "gpt-5-nano",
133
- input_token_cost: 0.05 / 1_000_000,
134
- output_token_cost: 0.4 / 1_000_000,
135
- model_provider_settings: { supports_temperature: false },
136
- }
137
179
  ]
138
180
 
139
181
  open_ai_responses_models = open_ai_models.dup.map.with_index do |model, _index|
@@ -150,7 +192,7 @@ module Raif
150
192
  # o1-mini is not supported by the OpenAI Responses API.
151
193
  open_ai_responses_models.delete_if{|model| model[:key] == :open_ai_o1_mini }
152
194
 
153
- # o1-pro and o3-pro are not supported by the OpenAI Completions API, but it is supported by the OpenAI Responses API.
195
+ # These models are not supported by the OpenAI Completions API, but are supported by the OpenAI Responses API.
154
196
  open_ai_responses_models << {
155
197
  key: :open_ai_responses_o1_pro,
156
198
  api_name: "o1-pro",
@@ -167,16 +209,107 @@ module Raif
167
209
  model_provider_settings: { supports_temperature: false },
168
210
  }
169
211
 
212
+ open_ai_responses_models << {
213
+ key: :open_ai_responses_gpt_5_pro,
214
+ api_name: "gpt-5-pro",
215
+ input_token_cost: 15.0 / 1_000_000,
216
+ output_token_cost: 120.0 / 1_000_000,
217
+ model_provider_settings: { supports_temperature: false },
218
+ }
219
+
220
+ open_ai_responses_models << {
221
+ key: :open_ai_responses_gpt_5_2_pro,
222
+ api_name: "gpt-5.2-pro",
223
+ input_token_cost: 21.0 / 1_000_000,
224
+ output_token_cost: 168.0 / 1_000_000,
225
+ model_provider_settings: { supports_temperature: false },
226
+ }
227
+
228
+ open_ai_responses_models << {
229
+ key: :open_ai_responses_gpt_5_4_pro,
230
+ api_name: "gpt-5.4-pro",
231
+ input_token_cost: 30.0 / 1_000_000,
232
+ output_token_cost: 180.0 / 1_000_000,
233
+ model_provider_settings: { supports_temperature: false, supports_structured_outputs: false },
234
+ }
235
+
170
236
  {
171
237
  Raif::Llms::OpenAiCompletions => open_ai_models,
172
238
  Raif::Llms::OpenAiResponses => open_ai_responses_models,
173
239
  Raif::Llms::Anthropic => [
174
240
  {
175
- key: :anthropic_claude_4_sonnet,
176
- api_name: "claude-sonnet-4-20250514",
241
+ key: :anthropic_claude_4_7_opus,
242
+ api_name: "claude-opus-4-7",
243
+ input_token_cost: 5.0 / 1_000_000,
244
+ output_token_cost: 25.0 / 1_000_000,
245
+ max_completion_tokens: 128_000,
246
+ model_provider_settings: { supports_temperature: false },
247
+ supported_provider_managed_tools: [
248
+ Raif::ModelTools::ProviderManaged::WebSearch,
249
+ Raif::ModelTools::ProviderManaged::CodeExecution
250
+ ]
251
+ },
252
+ {
253
+ key: :anthropic_claude_4_6_opus,
254
+ api_name: "claude-opus-4-6",
255
+ input_token_cost: 5.0 / 1_000_000,
256
+ output_token_cost: 25.0 / 1_000_000,
257
+ max_completion_tokens: 128_000,
258
+ supported_provider_managed_tools: [
259
+ Raif::ModelTools::ProviderManaged::WebSearch,
260
+ Raif::ModelTools::ProviderManaged::CodeExecution
261
+ ]
262
+ },
263
+ {
264
+ key: :anthropic_claude_4_6_sonnet,
265
+ api_name: "claude-sonnet-4-6",
177
266
  input_token_cost: 3.0 / 1_000_000,
178
267
  output_token_cost: 15.0 / 1_000_000,
179
- max_completion_tokens: 8192,
268
+ max_completion_tokens: 64_000,
269
+ supported_provider_managed_tools: [
270
+ Raif::ModelTools::ProviderManaged::WebSearch,
271
+ Raif::ModelTools::ProviderManaged::CodeExecution
272
+ ]
273
+ },
274
+ {
275
+ key: :anthropic_claude_4_5_opus,
276
+ api_name: "claude-opus-4-5",
277
+ input_token_cost: 5.0 / 1_000_000,
278
+ output_token_cost: 25.0 / 1_000_000,
279
+ max_completion_tokens: 64_000,
280
+ supported_provider_managed_tools: [
281
+ Raif::ModelTools::ProviderManaged::WebSearch,
282
+ Raif::ModelTools::ProviderManaged::CodeExecution
283
+ ]
284
+ },
285
+ {
286
+ key: :anthropic_claude_4_5_sonnet,
287
+ api_name: "claude-sonnet-4-5",
288
+ input_token_cost: 3.0 / 1_000_000,
289
+ output_token_cost: 15.0 / 1_000_000,
290
+ max_completion_tokens: 64_000,
291
+ supported_provider_managed_tools: [
292
+ Raif::ModelTools::ProviderManaged::WebSearch,
293
+ Raif::ModelTools::ProviderManaged::CodeExecution
294
+ ]
295
+ },
296
+ {
297
+ key: :anthropic_claude_4_5_haiku,
298
+ api_name: "claude-haiku-4-5",
299
+ input_token_cost: 1.0 / 1_000_000,
300
+ output_token_cost: 5.0 / 1_000_000,
301
+ max_completion_tokens: 64_000,
302
+ supported_provider_managed_tools: [
303
+ Raif::ModelTools::ProviderManaged::WebSearch,
304
+ Raif::ModelTools::ProviderManaged::CodeExecution
305
+ ]
306
+ },
307
+ {
308
+ key: :anthropic_claude_4_1_opus,
309
+ api_name: "claude-opus-4-1",
310
+ input_token_cost: 15.0 / 1_000_000,
311
+ output_token_cost: 75.0 / 1_000_000,
312
+ max_completion_tokens: 32_000,
180
313
  supported_provider_managed_tools: [
181
314
  Raif::ModelTools::ProviderManaged::WebSearch,
182
315
  Raif::ModelTools::ProviderManaged::CodeExecution
@@ -187,7 +320,18 @@ module Raif
187
320
  api_name: "claude-opus-4-20250514",
188
321
  input_token_cost: 15.0 / 1_000_000,
189
322
  output_token_cost: 75.0 / 1_000_000,
190
- max_completion_tokens: 8192,
323
+ max_completion_tokens: 32_000,
324
+ supported_provider_managed_tools: [
325
+ Raif::ModelTools::ProviderManaged::WebSearch,
326
+ Raif::ModelTools::ProviderManaged::CodeExecution
327
+ ]
328
+ },
329
+ {
330
+ key: :anthropic_claude_4_sonnet,
331
+ api_name: "claude-sonnet-4-20250514",
332
+ input_token_cost: 3.0 / 1_000_000,
333
+ output_token_cost: 15.0 / 1_000_000,
334
+ max_completion_tokens: 64_000,
191
335
  supported_provider_managed_tools: [
192
336
  Raif::ModelTools::ProviderManaged::WebSearch,
193
337
  Raif::ModelTools::ProviderManaged::CodeExecution
@@ -235,30 +379,79 @@ module Raif
235
379
  },
236
380
  ],
237
381
  Raif::Llms::Bedrock => [
382
+ {
383
+ key: :bedrock_claude_4_7_opus,
384
+ api_name: "anthropic.claude-opus-4-7",
385
+ input_token_cost: 0.005 / 1000,
386
+ output_token_cost: 0.025 / 1000,
387
+ max_completion_tokens: 128_000
388
+ },
389
+ {
390
+ key: :bedrock_claude_4_6_opus,
391
+ api_name: "anthropic.claude-opus-4-6-v1",
392
+ input_token_cost: 0.005 / 1000,
393
+ output_token_cost: 0.025 / 1000,
394
+ max_completion_tokens: 128_000
395
+ },
396
+ {
397
+ key: :bedrock_claude_4_6_sonnet,
398
+ api_name: "anthropic.claude-sonnet-4-6",
399
+ input_token_cost: 0.003 / 1000,
400
+ output_token_cost: 0.015 / 1000,
401
+ max_completion_tokens: 64_000
402
+ },
403
+ {
404
+ key: :bedrock_claude_4_5_opus,
405
+ api_name: "anthropic.claude-opus-4-5-20251101-v1:0",
406
+ input_token_cost: 0.005 / 1000,
407
+ output_token_cost: 0.025 / 1000,
408
+ max_completion_tokens: 64_000
409
+ },
410
+ {
411
+ key: :bedrock_claude_4_5_sonnet,
412
+ api_name: "anthropic.claude-sonnet-4-5-20250929-v1:0",
413
+ input_token_cost: 0.003 / 1000,
414
+ output_token_cost: 0.015 / 1000,
415
+ max_completion_tokens: 64_000
416
+ },
417
+ {
418
+ key: :bedrock_claude_4_5_haiku,
419
+ api_name: "anthropic.claude-haiku-4-5-20251001-v1:0",
420
+ input_token_cost: 0.001 / 1000,
421
+ output_token_cost: 0.005 / 1000,
422
+ max_completion_tokens: 64_000
423
+ },
424
+ {
425
+ key: :bedrock_claude_4_1_opus,
426
+ api_name: "anthropic.claude-opus-4-1-20250805-v1:0",
427
+ input_token_cost: 0.015 / 1000,
428
+ output_token_cost: 0.075 / 1000,
429
+ max_completion_tokens: 32_000
430
+ },
238
431
  {
239
432
  key: :bedrock_claude_4_sonnet,
240
433
  api_name: "anthropic.claude-sonnet-4-20250514-v1:0",
241
434
  input_token_cost: 0.003 / 1000,
242
435
  output_token_cost: 0.015 / 1000,
243
- max_completion_tokens: 8192
436
+ max_completion_tokens: 64_000
244
437
  },
245
438
  {
246
439
  key: :bedrock_claude_4_opus,
247
440
  api_name: "anthropic.claude-opus-4-20250514-v1:0",
248
441
  input_token_cost: 0.015 / 1000,
249
442
  output_token_cost: 0.075 / 1000,
250
- max_completion_tokens: 8192
443
+ max_completion_tokens: 32_000
251
444
  },
252
445
  {
253
- key: :bedrock_claude_3_5_sonnet,
254
- api_name: "anthropic.claude-3-5-sonnet-20241022-v2:0",
446
+ key: :bedrock_claude_3_7_sonnet,
447
+ api_name: "anthropic.claude-3-7-sonnet-20250219-v1:0",
255
448
  input_token_cost: 0.003 / 1000,
256
449
  output_token_cost: 0.015 / 1000,
257
450
  max_completion_tokens: 8192
258
451
  },
259
452
  {
260
- key: :bedrock_claude_3_7_sonnet,
261
- api_name: "anthropic.claude-3-7-sonnet-20250219-v1:0",
453
+ key: :bedrock_claude_3_5_sonnet,
454
+ api_name: "anthropic.claude-3-5-sonnet-20241022-v2:0",
262
455
  input_token_cost: 0.003 / 1000,
263
456
  output_token_cost: 0.015 / 1000,
264
457
  max_completion_tokens: 8192
@@ -297,6 +490,34 @@ module Raif
297
490
  input_token_cost: 0.0002625 / 1000,
298
491
  output_token_cost: 0.0042 / 1000,
299
492
  max_completion_tokens: 4096
493
+ },
494
+ {
495
+ key: :bedrock_deepseek_v3_2,
496
+ api_name: "deepseek.v3.2",
497
+ input_token_cost: 0.62 / 1_000_000,
498
+ output_token_cost: 1.85 / 1_000_000,
499
+ max_completion_tokens: 32_768
500
+ },
501
+ {
502
+ key: :bedrock_deepseek_r1,
503
+ api_name: "deepseek.r1-v1:0",
504
+ input_token_cost: 0.00135 / 1_000,
505
+ output_token_cost: 0.0054 / 1_000,
506
+ max_completion_tokens: 32_768
507
+ },
508
+ {
509
+ key: :bedrock_gpt_oss_120b,
510
+ api_name: "openai.gpt-oss-120b-1:0",
511
+ input_token_cost: 0.15 / 1_000_000,
512
+ output_token_cost: 0.60 / 1_000_000,
513
+ max_completion_tokens: 32_768
514
+ },
515
+ {
516
+ key: :bedrock_gpt_oss_20b,
517
+ api_name: "openai.gpt-oss-20b-1:0",
518
+ input_token_cost: 0.07 / 1_000_000,
519
+ output_token_cost: 0.30 / 1_000_000,
520
+ max_completion_tokens: 32_768
300
521
  }
301
522
  ],
302
523
  Raif::Llms::OpenRouter => [
@@ -306,6 +527,84 @@ module Raif
306
527
  input_token_cost: 3.0 / 1_000_000,
307
528
  output_token_cost: 15.0 / 1_000_000,
308
529
  },
530
+ {
531
+ key: :open_router_deepseek_chat_v3,
532
+ api_name: "deepseek/deepseek-chat-v3-0324",
533
+ input_token_cost: 0.20 / 1_000_000,
534
+ output_token_cost: 0.77 / 1_000_000,
535
+ },
536
+ {
537
+ key: :open_router_deepseek_v3_1,
538
+ api_name: "deepseek/deepseek-chat-v3.1",
539
+ input_token_cost: 0.25 / 1_000_000,
540
+ output_token_cost: 1.0 / 1_000_000,
541
+ },
542
+ {
543
+ key: :open_router_deepseek_v3_2,
544
+ api_name: "deepseek/deepseek-v3.2",
545
+ input_token_cost: 0.26 / 1_000_000,
546
+ output_token_cost: 0.38 / 1_000_000,
547
+ },
548
+ {
549
+ key: :open_router_gemini_2_0_flash,
550
+ api_name: "google/gemini-2.0-flash-001",
551
+ input_token_cost: 0.1 / 1_000_000,
552
+ output_token_cost: 0.4 / 1_000_000,
553
+ },
554
+ {
555
+ key: :open_router_gemini_2_5_flash,
556
+ api_name: "google/gemini-2.5-flash",
557
+ input_token_cost: 0.3 / 1_000_000,
558
+ output_token_cost: 2.5 / 1_000_000,
559
+ },
560
+ {
561
+ key: :open_router_gemini_2_5_pro,
562
+ api_name: "google/gemini-2.5-pro",
563
+ input_token_cost: 1.25 / 1_000_000,
564
+ output_token_cost: 10.0 / 1_000_000,
565
+ },
566
+ {
567
+ key: :open_router_gemini_3_pro_preview,
568
+ api_name: "google/gemini-3-pro-preview",
569
+ input_token_cost: 2.0 / 1_000_000,
570
+ output_token_cost: 12.0 / 1_000_000,
571
+ },
572
+ {
573
+ key: :open_router_gemini_3_1_pro_preview,
574
+ api_name: "google/gemini-3.1-pro-preview",
575
+ input_token_cost: 2.0 / 1_000_000,
576
+ output_token_cost: 12.0 / 1_000_000,
577
+ },
578
+ {
579
+ key: :open_router_gemini_3_1_flash_lite_preview,
580
+ api_name: "google/gemini-3.1-flash-lite-preview",
581
+ input_token_cost: 0.25 / 1_000_000,
582
+ output_token_cost: 1.5 / 1_000_000,
583
+ },
584
+ {
585
+ key: :open_router_grok_4,
586
+ api_name: "x-ai/grok-4",
587
+ input_token_cost: 3.0 / 1_000_000,
588
+ output_token_cost: 15.0 / 1_000_000,
589
+ },
590
+ {
591
+ key: :open_router_grok_4_1_fast,
592
+ api_name: "x-ai/grok-4.1-fast",
593
+ input_token_cost: 0.2 / 1_000_000,
594
+ output_token_cost: 0.5 / 1_000_000,
595
+ },
596
+ {
597
+ key: :open_router_kimi_k2_thinking,
598
+ api_name: "moonshotai/kimi-k2-thinking",
599
+ input_token_cost: 0.45 / 1_000_000,
600
+ output_token_cost: 2.35 / 1_000_000,
601
+ },
602
+ {
603
+ key: :open_router_kimi_k2_5,
604
+ api_name: "moonshotai/kimi-k2.5",
605
+ input_token_cost: 0.45 / 1_000_000,
606
+ output_token_cost: 2.20 / 1_000_000,
607
+ },
309
608
  {
310
609
  key: :open_router_llama_3_3_70b_instruct,
311
610
  api_name: "meta-llama/llama-3.3-70b-instruct",
@@ -331,16 +630,34 @@ module Raif
331
630
  output_token_cost: 0.30 / 1_000_000,
332
631
  },
333
632
  {
334
- key: :open_router_gemini_2_0_flash,
335
- api_name: "google/gemini-2.0-flash-001",
336
- input_token_cost: 0.1 / 1_000_000,
337
- output_token_cost: 0.4 / 1_000_000,
633
+ key: :open_router_minimax_m2,
634
+ api_name: "minimax/minimax-m2",
635
+ input_token_cost: 0.255 / 1_000_000,
636
+ output_token_cost: 1.02 / 1_000_000,
338
637
  },
339
638
  {
340
- key: :open_router_deepseek_chat_v3,
341
- api_name: "deepseek/deepseek-chat-v3-0324",
639
+ key: :open_router_minimax_m2_1,
640
+ api_name: "minimax/minimax-m2.1",
342
641
  input_token_cost: 0.27 / 1_000_000,
343
- output_token_cost: 1.1 / 1_000_000,
642
+ output_token_cost: 0.95 / 1_000_000,
643
+ },
644
+ {
645
+ key: :open_router_minimax_m2_5,
646
+ api_name: "minimax/minimax-m2.5",
647
+ input_token_cost: 0.27 / 1_000_000,
648
+ output_token_cost: 0.95 / 1_000_000,
649
+ },
650
+ {
651
+ key: :open_router_mistral_large_3_2512,
652
+ api_name: "mistralai/mistral-large-2512",
653
+ input_token_cost: 0.50 / 1_000_000,
654
+ output_token_cost: 1.5 / 1_000_000,
655
+ },
656
+ {
657
+ key: :open_router_mistral_small_3_2_24b,
658
+ api_name: "mistralai/mistral-small-3.2-24b-instruct",
659
+ input_token_cost: 0.06 / 1_000_000,
660
+ output_token_cost: 0.18 / 1_000_000,
344
661
  },
345
662
  {
346
663
  key: :open_router_open_ai_gpt_oss_120b,
@@ -353,7 +670,81 @@ module Raif
353
670
  api_name: "gpt-oss-20b",
354
671
  input_token_cost: 0.05 / 1_000_000,
355
672
  output_token_cost: 0.2 / 1_000_000,
356
- }
673
+ },
674
+ {
675
+ key: :open_router_grok_4_20,
676
+ api_name: "x-ai/grok-4.20",
677
+ input_token_cost: 2.0 / 1_000_000,
678
+ output_token_cost: 6.0 / 1_000_000,
679
+ },
680
+ {
681
+ key: :open_router_google_gemma_4_31b_it,
682
+ api_name: "google/gemma-4-31b-it",
683
+ input_token_cost: 0.14 / 1_000_000,
684
+ output_token_cost: 0.40 / 1_000_000,
685
+ },
686
+ ],
687
+ Raif::Llms::Google => [
688
+ {
689
+ key: :google_gemini_3_1_pro,
690
+ api_name: "gemini-3.1-pro-preview",
691
+ input_token_cost: 2.0 / 1_000_000,
692
+ output_token_cost: 12.0 / 1_000_000,
693
+ supported_provider_managed_tools: [
694
+ Raif::ModelTools::ProviderManaged::WebSearch,
695
+ Raif::ModelTools::ProviderManaged::CodeExecution
696
+ ]
697
+ },
698
+ {
699
+ key: :google_gemini_3_1_flash_lite,
700
+ api_name: "gemini-3.1-flash-lite-preview",
701
+ input_token_cost: 0.25 / 1_000_000,
702
+ output_token_cost: 1.5 / 1_000_000,
703
+ supported_provider_managed_tools: [
704
+ Raif::ModelTools::ProviderManaged::WebSearch,
705
+ Raif::ModelTools::ProviderManaged::CodeExecution
706
+ ]
707
+ },
708
+ {
709
+ key: :google_gemini_3_0_pro,
710
+ api_name: "gemini-3-pro-preview",
711
+ input_token_cost: 2.0 / 1_000_000,
712
+ output_token_cost: 12.0 / 1_000_000,
713
+ supported_provider_managed_tools: [
714
+ Raif::ModelTools::ProviderManaged::WebSearch,
715
+ Raif::ModelTools::ProviderManaged::CodeExecution
716
+ ]
717
+ },
718
+ {
719
+ key: :google_gemini_3_0_flash,
720
+ api_name: "gemini-3-flash-preview",
721
+ input_token_cost: 0.5 / 1_000_000,
722
+ output_token_cost: 3.0 / 1_000_000,
723
+ supported_provider_managed_tools: [
724
+ Raif::ModelTools::ProviderManaged::WebSearch,
725
+ Raif::ModelTools::ProviderManaged::CodeExecution
726
+ ]
727
+ },
728
+ {
729
+ key: :google_gemini_2_5_pro,
730
+ api_name: "gemini-2.5-pro-preview-06-05",
731
+ input_token_cost: 1.25 / 1_000_000,
732
+ output_token_cost: 10.0 / 1_000_000,
733
+ supported_provider_managed_tools: [
734
+ Raif::ModelTools::ProviderManaged::WebSearch,
735
+ Raif::ModelTools::ProviderManaged::CodeExecution
736
+ ]
737
+ },
738
+ {
739
+ key: :google_gemini_2_5_flash,
740
+ api_name: "gemini-2.5-flash",
741
+ input_token_cost: 0.3 / 1_000_000,
742
+ output_token_cost: 2.5 / 1_000_000,
743
+ supported_provider_managed_tools: [
744
+ Raif::ModelTools::ProviderManaged::WebSearch,
745
+ Raif::ModelTools::ProviderManaged::CodeExecution
746
+ ]
747
+ },
357
748
  ]
358
749
  }
359
750
  end