llm_cost_tracker 0.15.1 → 0.15.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 5c832e2df79470701d964d7c28c67122c60b5e51f8d81aadfe09f396601b14ef
-  data.tar.gz: c26760c25db1e3ff3bf90f89a7141c72b83ad7ef5b9acbc76c19ba5468f65b6f
+  metadata.gz: 380c06ddfdda1e5df94061a377dc703fd1f52c4b5ab92ab1f786ef1fce591be1
+  data.tar.gz: e58224c8eb4d538faea70fe867d33be5000611a35ec58ab33feb72d39fca9ef6
 SHA512:
-  metadata.gz: cd7c26e49a779e97a3f054a41d09aa2a114104ede4ebe4211f08adb5941ecd8f5dcaa325af7ff4aa1b17ca582a2422f50566a6aa8eb691b7cffe2a91e1c79f28
-  data.tar.gz: 4236c7bbdf92ae68e8ce3f33763795211bb5e97bdf2c7884d18dee6bed7c640abd7796c506a0ce8406b78ecf15b7eaa9c26fa8b02d7a11c8da9c76bfd27e6e38
+  metadata.gz: 2cd2b3c33e60e7638c58044220f2e43336be78f2cd085d253cdf99badb9af8586f411d1ced9e332a7255b81f08267407a52e7610b5dedc33fc2623cbca681a34
+  data.tar.gz: 70c17315f8a243cc59ebaf036f480151320a4ecb3adc53d152f5f15bfe50b51d81fdc0722e784d026d540e94d64d9a64443cc6f8cf51bcaf5585093b80ee049a

data/CHANGELOG.md

@@ -2,6 +2,17 @@
 
 Format: [Keep a Changelog](https://keepachangelog.com/en/1.1.0/). Versioning: [SemVer](https://semver.org/spec/v2.0.0.html).
 
+## [0.15.2] - 2026-10-29
+
+### Changed
+
+- Bundled prices add OpenAI GPT-6.1 Terra.
+
+### Fixed
+
+- OpenAI Decisions API calls through the official openai gem and Faraday are recorded.
+- OpenAI Decisions output tokens are not billed, RubyLLM 2.x judgments included, as OpenAI prices them.
+
 ## [0.15.1] - 2026-10-26
 
 ### Changed

data/lib/llm_cost_tracker/integrations/openai.rb

@@ -47,6 +47,7 @@ module LlmCostTracker
             patch_target("OpenAI::Resources::Audio::Translations", with: TranslationsPatch, optional: true),
             patch_target("OpenAI::Resources::Audio::Speech", with: SpeechPatch, optional: true),
             patch_target("OpenAI::Resources::Moderations", with: ModerationsPatch, optional: true),
+            patch_target("OpenAI::Resources::Decisions", with: CreatePatch, optional: true, since: "0.101.0"),
             patch_target("OpenAI::Resources::Batches", with: BatchesPatch, optional: true)
           ]
         end

data/lib/llm_cost_tracker/prices.json

@@ -29,6 +29,7 @@
       "openrouter/arcee-ai/coder-large": "2026-10-05",
       "openrouter/arcee-ai/maestro-reasoning": "2026-10-05",
       "openrouter/arcee-ai/spotlight": "2026-10-05",
+      "openrouter/arcee-ai/trinity-large-thinking": "2026-10-27",
       "openrouter/arcee-ai/trinity-mini": "2026-10-05",
       "openrouter/arcee-ai/virtuoso-large": "2026-10-05",
       "openrouter/baidu/ernie-4.5-21b-a3b": "2026-10-05",
@@ -55,6 +56,7 @@
       "openrouter/google/gemini-2.5-pro:batch": "2026-10-21",
       "openrouter/google/gemma-3n-e4b-it": "2026-10-05",
       "openrouter/google/gemma-4-31b-it:batch": "2026-10-05",
+      "openrouter/ibm-granite/granite-4.0-h-micro": "2026-10-27",
       "openrouter/ibm-granite/granite-4.1-8b": "2026-10-05",
       "openrouter/inception/mercury-2.5-preview": "2026-10-05",
       "openrouter/inclusionai/ling-2.6-1t": "2026-10-05",
@@ -76,11 +78,16 @@
       "openrouter/microsoft/phi-4-mini-instruct": "2026-10-05",
       "openrouter/microsoft/wizardlm-2-8x22b": "2026-10-19",
       "openrouter/minimax/minimax-m3:batch": "2026-10-05",
+      "openrouter/mistralai/codestral-2508:batch": "2026-10-27",
       "openrouter/mistralai/devstral-medium": "2026-10-05",
       "openrouter/mistralai/devstral-small": "2026-10-05",
       "openrouter/mistralai/ministral-8b": "2026-10-05",
+      "openrouter/mistralai/ministral-8b-2512:batch": "2026-10-27",
       "openrouter/mistralai/mistral-7b-instruct-v0.1": "2026-10-05",
       "openrouter/mistralai/mistral-large-2411": "2026-10-05",
+      "openrouter/mistralai/mistral-large-2512:batch": "2026-10-27",
+      "openrouter/mistralai/mistral-medium-3.1:batch": "2026-10-27",
+      "openrouter/mistralai/mistral-small-2603:batch": "2026-10-27",
       "openrouter/mistralai/pixtral-large-2411": "2026-10-05",
       "openrouter/moonshotai/kimi-k2.7-code:batch": "2026-10-05",
       "openrouter/nex-agi/deepseek-v3.1-nex-n1": "2026-10-05",
@@ -198,7 +205,7 @@
       "https://docs.x.ai/developers/pricing.md"
     ],
     "unit": "1M tokens",
-    "updated_at": "2026-10-26"
+    "updated_at": "2026-10-28"
   },
   "models": {
     "anthropic/claude-3-5-haiku": {
@@ -3774,6 +3781,109 @@
       "ultrafast_input": 12.0,
       "ultrafast_output": 60.0
     },
+    "openai/gpt-6.1-terra": {
+      "_context_price_threshold_tokens": 272000,
+      "above_context_batch_cache_read_input": 0.05,
+      "above_context_batch_cache_write_extended_input": 2.5,
+      "above_context_batch_cache_write_input": 1.25,
+      "above_context_batch_data_residency_cache_read_input": 0.055,
+      "above_context_batch_data_residency_cache_write_extended_input": 2.75,
+      "above_context_batch_data_residency_cache_write_input": 1.375,
+      "above_context_batch_data_residency_input": 1.1,
+      "above_context_batch_data_residency_output": 4.125,
+      "above_context_batch_input": 1.0,
+      "above_context_batch_output": 3.75,
+      "above_context_cache_read_input": 0.1,
+      "above_context_cache_write_extended_input": 5.0,
+      "above_context_cache_write_input": 2.5,
+      "above_context_data_residency_cache_read_input": 0.11,
+      "above_context_data_residency_cache_write_extended_input": 5.5,
+      "above_context_data_residency_cache_write_input": 2.75,
+      "above_context_data_residency_input": 2.2,
+      "above_context_data_residency_output": 8.25,
+      "above_context_fast_cache_read_input": 0.2,
+      "above_context_fast_cache_write_extended_input": 10.0,
+      "above_context_fast_cache_write_input": 5.0,
+      "above_context_fast_data_residency_cache_read_input": 0.22,
+      "above_context_fast_data_residency_cache_write_extended_input": 11.0,
+      "above_context_fast_data_residency_cache_write_input": 5.5,
+      "above_context_fast_data_residency_input": 4.4,
+      "above_context_fast_data_residency_output": 16.5,
+      "above_context_fast_input": 4.0,
+      "above_context_fast_output": 15.0,
+      "above_context_flex_cache_read_input": 0.05,
+      "above_context_flex_cache_write_extended_input": 2.5,
+      "above_context_flex_cache_write_input": 1.25,
+      "above_context_flex_data_residency_cache_read_input": 0.055,
+      "above_context_flex_data_residency_cache_write_extended_input": 2.75,
+      "above_context_flex_data_residency_cache_write_input": 1.375,
+      "above_context_flex_data_residency_input": 1.1,
+      "above_context_flex_data_residency_output": 4.125,
+      "above_context_flex_input": 1.0,
+      "above_context_flex_output": 3.75,
+      "above_context_input": 2.0,
+      "above_context_output": 7.5,
+      "above_context_priority_cache_read_input": 0.2,
+      "above_context_priority_cache_write_extended_input": 10.0,
+      "above_context_priority_cache_write_input": 5.0,
+      "above_context_priority_data_residency_cache_read_input": 0.22,
+      "above_context_priority_data_residency_cache_write_extended_input": 11.0,
+      "above_context_priority_data_residency_cache_write_input": 5.5,
+      "above_context_priority_data_residency_input": 4.4,
+      "above_context_priority_data_residency_output": 16.5,
+      "above_context_priority_input": 4.0,
+      "above_context_priority_output": 15.0,
+      "batch_cache_read_input": 0.025,
+      "batch_cache_write_extended_input": 1.25,
+      "batch_cache_write_input": 0.625,
+      "batch_data_residency_cache_read_input": 0.0275,
+      "batch_data_residency_cache_write_extended_input": 1.375,
+      "batch_data_residency_cache_write_input": 0.6875,
+      "batch_data_residency_input": 0.55,
+      "batch_data_residency_output": 2.75,
+      "batch_input": 0.5,
+      "batch_output": 2.5,
+      "cache_read_input": 0.05,
+      "cache_write_extended_input": 2.5,
+      "cache_write_input": 1.25,
+      "data_residency_cache_read_input": 0.055,
+      "data_residency_cache_write_extended_input": 2.75,
+      "data_residency_cache_write_input": 1.375,
+      "data_residency_input": 1.1,
+      "data_residency_output": 5.5,
+      "fast_cache_read_input": 0.1,
+      "fast_cache_write_extended_input": 5.0,
+      "fast_cache_write_input": 2.5,
+      "fast_data_residency_cache_read_input": 0.11,
+      "fast_data_residency_cache_write_extended_input": 5.5,
+      "fast_data_residency_cache_write_input": 2.75,
+      "fast_data_residency_input": 2.2,
+      "fast_data_residency_output": 11.0,
+      "fast_input": 2.0,
+      "fast_output": 10.0,
+      "flex_cache_read_input": 0.025,
+      "flex_cache_write_extended_input": 1.25,
+      "flex_cache_write_input": 0.625,
+      "flex_data_residency_cache_read_input": 0.0275,
+      "flex_data_residency_cache_write_extended_input": 1.375,
+      "flex_data_residency_cache_write_input": 0.6875,
+      "flex_data_residency_input": 0.55,
+      "flex_data_residency_output": 2.75,
+      "flex_input": 0.5,
+      "flex_output": 2.5,
+      "input": 1.0,
+      "output": 5.0,
+      "priority_cache_read_input": 0.1,
+      "priority_cache_write_extended_input": 5.0,
+      "priority_cache_write_input": 2.5,
+      "priority_data_residency_cache_read_input": 0.11,
+      "priority_data_residency_cache_write_extended_input": 5.5,
+      "priority_data_residency_cache_write_input": 2.75,
+      "priority_data_residency_input": 2.2,
+      "priority_data_residency_output": 11.0,
+      "priority_input": 2.0,
+      "priority_output": 10.0
+    },
     "openai/gpt-audio": {
       "audio_input": 32.0,
       "audio_output": 64.0,
@@ -4478,9 +4588,9 @@
       "output": 0.182
     },
     "openrouter/deepseek/deepseek-v4-flash-0731": {
-      "cache_read_input": 0.028,
-      "input": 0.0422,
-      "output": 0.1694
+      "cache_read_input": 0.0238,
+      "input": 0.119,
+      "output": 0.238
     },
     "openrouter/deepseek/deepseek-v4-flash-0731:batch": {
       "cache_read_input": 0.0035,
@@ -4498,9 +4608,9 @@
       "output": 0.33
     },
     "openrouter/deepseek/deepseek-v4-pro": {
-      "cache_read_input": 0.0174,
-      "input": 0.2088,
-      "output": 0.4176
+      "cache_read_input": 0.079605,
+      "input": 0.95526,
+      "output": 1.91052
     },
     "openrouter/deepseek/deepseek-v4-pro-0813": {
       "cache_read_input": 0.044,
@@ -4513,9 +4623,9 @@
       "output": 1.98
     },
     "openrouter/deepseek/deepseek-v4.1-flash": {
-      "cache_read_input": 0.015,
-      "input": 0.02,
-      "output": 1.0
+      "cache_read_input": 0.0042,
+      "input": 0.14,
+      "output": 0.42
     },
     "openrouter/deepseek/deepseek-v4.1-flash-1017": {
       "cache_read_input": 0.006,
@@ -4831,18 +4941,17 @@
       "output": 0.35
     },
     "openrouter/google/gemma-3-4b-it": {
-      "cache_read_input": 0.02,
-      "input": 0.04,
-      "output": 0.05
+      "input": 0.05,
+      "output": 0.1
     },
     "openrouter/google/gemma-3n-e4b-it": {
       "input": 0.06,
       "output": 0.12
     },
     "openrouter/google/gemma-4-26b-a4b-it": {
-      "cache_read_input": 0.0375,
-      "input": 0.0675,
-      "output": 0.225
+      "cache_read_input": 0.05,
+      "input": 0.09,
+      "output": 0.3
     },
     "openrouter/google/gemma-4-31b-it": {
       "input": 0.2,
@@ -4999,9 +5108,8 @@
       "output": 0.4
     },
     "openrouter/meta-llama/llama-3.1-8b-instruct": {
-      "cache_read_input": 0.025,
-      "input": 0.05,
-      "output": 0.08
+      "input": 0.02,
+      "output": 0.04
     },
     "openrouter/meta-llama/llama-3.2-11b-vision-instruct": {
       "input": 0.345,
@@ -5327,8 +5435,8 @@
     },
     "openrouter/moonshotai/kimi-k2.7-code": {
       "cache_read_input": 0.15,
-      "input": 0.6712,
-      "output": 3.35
+      "input": 0.56,
+      "output": 3.1
     },
     "openrouter/moonshotai/kimi-k2.7-code:batch": {
       "cache_read_input": 0.19,
@@ -5336,9 +5444,9 @@
       "output": 4.0
     },
     "openrouter/moonshotai/kimi-k3": {
-      "cache_read_input": 0.21,
-      "input": 0.239,
-      "output": 10.8
+      "cache_read_input": 0.169,
+      "input": 0.179,
+      "output": 20.0
     },
     "openrouter/moonshotai/kimi-k3.1-pro": {
       "cache_read_input": 0.2,
@@ -5395,6 +5503,7 @@
       "output": 0.7
     },
     "openrouter/nousresearch/hermes-4-405b": {
+      "cache_read_input": 0.2,
       "input": 1.0,
       "output": 3.0
     },
@@ -5407,6 +5516,7 @@
       "output": 0.4
     },
     "openrouter/nvidia/nemotron-3-nano-30b-a3b": {
+      "cache_read_input": 0.012,
       "input": 0.06,
       "output": 0.24
     },
@@ -5503,7 +5613,7 @@
       "output": 0.8
     },
     "openrouter/openai/gpt-4.1-nano": {
-      "cache_read_input": 0.025,
+      "cache_read_input": 0.03,
       "input": 0.1,
       "output": 0.4
     },
@@ -5925,6 +6035,30 @@
       "input": 1.0,
       "output": 5.0
     },
+    "openrouter/openai/gpt-6.1-terra": {
+      "cache_read_input": 0.05,
+      "cache_write_input": 1.25,
+      "input": 1.0,
+      "output": 5.0
+    },
+    "openrouter/openai/gpt-6.1-terra-pro": {
+      "cache_read_input": 0.05,
+      "cache_write_input": 1.25,
+      "input": 1.0,
+      "output": 5.0
+    },
+    "openrouter/openai/gpt-6.1-terra-pro:batch": {
+      "cache_read_input": 0.025,
+      "cache_write_input": 0.625,
+      "input": 0.5,
+      "output": 2.5
+    },
+    "openrouter/openai/gpt-6.1-terra:batch": {
+      "cache_read_input": 0.025,
+      "cache_write_input": 0.625,
+      "input": 0.5,
+      "output": 2.5
+    },
     "openrouter/openai/gpt-audio": {
       "audio_input": 32.0,
       "input": 2.5,
@@ -6106,6 +6240,11 @@
       "input": 0.075,
       "output": 0.5
     },
+    "openrouter/prism-ml/ternary-bonsai-2-9b": {
+      "cache_read_input": 0.04,
+      "input": 0.08,
+      "output": 0.13
+    },
     "openrouter/qwen/qwen-2.5-72b-instruct": {
       "input": 0.36,
       "output": 0.4
@@ -6158,7 +6297,8 @@
       "output": 0.5
     },
     "openrouter/qwen/qwen3-30b-a3b-instruct-2507": {
-      "input": 0.09,
+      "cache_read_input": 0.1,
+      "input": 0.1,
       "output": 0.3
     },
     "openrouter/qwen/qwen3-30b-a3b-thinking-2507": {
@@ -6248,13 +6388,12 @@
       "output": 2.1
     },
     "openrouter/qwen/qwen3.5-122b-a10b": {
-      "input": 0.4,
-      "output": 3.2
+      "input": 0.26,
+      "output": 2.08
     },
     "openrouter/qwen/qwen3.5-27b": {
-      "cache_read_input": 0.27,
-      "input": 0.27,
-      "output": 2.16
+      "input": 0.26,
+      "output": 2.6
     },
     "openrouter/qwen/qwen3.5-35b-a3b": {
       "cache_read_input": 0.05,
@@ -6288,14 +6427,13 @@
       "output": 1.8
     },
     "openrouter/qwen/qwen3.6-27b": {
-      "cache_read_input": 0.025,
-      "input": 0.25,
-      "output": 2.0
+      "input": 0.45,
+      "output": 2.7
     },
     "openrouter/qwen/qwen3.6-35b-a3b": {
-      "cache_read_input": 0.05,
+      "cache_read_input": 0.025,
       "input": 0.1,
-      "output": 0.95
+      "output": 0.9
     },
     "openrouter/qwen/qwen3.6-flash": {
       "cache_write_input": 0.234375,
@@ -6345,9 +6483,9 @@
       "output": 6.0
     },
     "openrouter/qwen/qwen3.8-27b": {
-      "cache_read_input": 0.09,
-      "input": 0.2,
-      "output": 2.5
+      "cache_read_input": 0.0039,
+      "input": 0.039,
+      "output": 1.25
     },
     "openrouter/qwen/qwen3.8-flash": {
       "cache_read_input": 0.016,
@@ -6628,14 +6766,14 @@
       "output": 3.0
     },
     "openrouter/xiaomi/mimo-v2.5": {
-      "cache_read_input": 0.00255,
-      "input": 0.119,
-      "output": 0.238
+      "cache_read_input": 0.08,
+      "input": 0.4,
+      "output": 2.0
     },
     "openrouter/xiaomi/mimo-v2.5-pro": {
-      "cache_read_input": 0.0028,
-      "input": 0.3045,
-      "output": 0.609
+      "cache_read_input": 0.096,
+      "input": 0.48,
+      "output": 1.8
     },
     "openrouter/xiaomi/mimo-v2.6-flash": {
       "cache_read_input": 0.0028,
@@ -6706,8 +6844,8 @@
       "output": 4.4
     },
     "openrouter/z-ai/glm-5.2": {
-      "cache_read_input": 0.033,
-      "input": 0.0341,
+      "cache_read_input": 0.013,
+      "input": 0.014,
       "output": 7.0
     },
     "openrouter/z-ai/glm-5.2:batch": {
@@ -6716,8 +6854,8 @@
       "output": 2.2
     },
     "openrouter/z-ai/glm-5.3": {
-      "cache_read_input": 0.078,
-      "input": 0.079,
+      "cache_read_input": 0.0689,
+      "input": 0.0699,
       "output": 1.4
     },
     "openrouter/z-ai/glm-5.3-flash": {

data/lib/llm_cost_tracker/providers/openai/parser.rb

@@ -8,7 +8,7 @@ module LlmCostTracker
         include StreamParser
 
         TRACKED_ENDPOINTS = %w[
-          chat/completions completions embeddings moderations responses responses/compact
+          chat/completions completions decisions embeddings moderations responses responses/compact
           audio/transcriptions audio/translations audio/speech
           images/generations images/edits images/variations
         ].freeze

data/lib/llm_cost_tracker/providers/openai/response_parser.rb

@@ -14,6 +14,7 @@ module LlmCostTracker
         PENDING_STATUSES = %w[queued in_progress].freeze
         RETRIEVE_PATH_PATTERN = %r{/(?:responses|agent)/resp_}
         TRANSCRIPTION_PATH_PATTERN = %r{/audio/(?:transcriptions|translations)\z}
+        UNBILLED_DECISION_USAGE = %i[output_tokens output_tokens_details total_tokens].freeze
 
         class << self
           def included(base) = base.include(StreamParser)
@@ -29,7 +30,7 @@ module LlmCostTracker
           end
 
           def event_from_response(response:, request:, provider:, host:, usage_source:, pricing_mode: nil)
-            usage = response["usage"]&.deep_symbolize_keys
+            usage = billed_usage(response, provider)
             return nil if usage.nil? || pending?(response)
 
             model = response["model"] || request["model"]
@@ -73,6 +74,15 @@ module LlmCostTracker
           def pending?(response)
             PENDING_STATUSES.include?(response["status"].to_s)
           end
+
+          def billed_usage(response, provider)
+            usage = response["usage"]&.deep_symbolize_keys
+            return usage unless usage && decision?(response, provider)
+
+            usage.except(*UNBILLED_DECISION_USAGE)
+          end
+
+          def decision?(response, provider) = provider.to_s == "openai" && response.key?("answers")
         end
 
         def parse(request_url:, request_body:, response_status:, response_body:, **)

data/lib/llm_cost_tracker/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module LlmCostTracker
-  VERSION = "0.15.1"
+  VERSION = "0.15.2"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: llm_cost_tracker
 version: !ruby/object:Gem::Version
-  version: 0.15.1
+  version: 0.15.2
 platform: ruby
 authors:
 - Sergii Khomenko