ruby_llm 1.14.1 → 1.16.0

data/lib/tasks/models.rake

@@ -329,25 +329,26 @@ end
 
 def standard_pricing_display(model)
   pricing_data = model.pricing.to_h[:text_tokens]&.dig(:standard) || {}
+  parts = [
+    pricing_part(pricing_data, :input_per_million, 'In'),
+    pricing_part(pricing_data, :output_per_million, 'Out'),
+    pricing_part(pricing_data, %i[cache_read_input_per_million cached_input_per_million], 'Cache Read'),
+    pricing_part(pricing_data, %i[cache_write_input_per_million cache_creation_input_per_million], 'Cache Write')
+  ].compact
 
-  if pricing_data.any?
-    parts = []
+  return parts.join(', ') if parts.any?
 
-    parts << "In: $#{format('%.2f', pricing_data[:input_per_million])}" if pricing_data[:input_per_million]
-
-    parts << "Out: $#{format('%.2f', pricing_data[:output_per_million])}" if pricing_data[:output_per_million]
-
-    if pricing_data[:cached_input_per_million]
-      parts << "Cache: $#{format('%.2f', pricing_data[:cached_input_per_million])}"
-    end
+  '-'
+end
 
-    return parts.join(', ') if parts.any?
-  end
+def pricing_part(pricing_data, key, label)
+  key = Array(key).find { |candidate| pricing_data[candidate] }
+  return unless key
 
-  '-'
+  "#{label}: $#{format('%.2f', pricing_data[key])}"
 end
 
-def generate_aliases # rubocop:disable Metrics/PerceivedComplexity
+def generate_aliases # rubocop:disable Metrics/AbcSize, Metrics/CyclomaticComplexity, Metrics/PerceivedComplexity
   models = Hash.new { |h, k| h[k] = [] }
 
   RubyLLM.models.all.each do |model|
@@ -466,12 +467,40 @@ def generate_aliases # rubocop:disable Metrics/PerceivedComplexity
     }
   end
 
+  add_xai_aliases(aliases, models['xai'])
+
   sorted_aliases = aliases.sort.to_h
   File.write(RubyLLM::Aliases.aliases_file, JSON.pretty_generate(sorted_aliases))
 
   puts "Generated #{sorted_aliases.size} aliases"
 end
 
+def add_xai_aliases(aliases, xai_models)
+  return unless xai_models.include?('grok-4.3')
+
+  %w[
+    grok-latest
+    grok-3
+    grok-3-latest
+    grok-3-mini
+    grok-3-mini-latest
+    grok-4
+    grok-4-latest
+    grok-4-fast
+    grok-4-fast-reasoning
+    grok-4-fast-reasoning-latest
+    grok-4-fast-non-reasoning
+    grok-4-fast-non-reasoning-latest
+    grok-4-1-fast
+    grok-4-1-fast-reasoning
+    grok-4-1-fast-reasoning-latest
+    grok-4-1-fast-non-reasoning
+    grok-4-1-fast-non-reasoning-latest
+  ].each do |alias_key|
+    aliases[alias_key] ||= { 'xai' => 'grok-4.3' }
+  end
+end
+
 def group_anthropic_models_by_base_name(anthropic_models)
   grouped = Hash.new { |h, k| h[k] = [] }
 
@@ -518,12 +547,12 @@ def find_best_bedrock_model(anthropic_model, bedrock_models) # rubocop:disable M
                  when 'claude-instant-v1', 'claude-instant'
                    'claude-instant'
                  else
-                   extract_base_name(anthropic_model)
+                   anthropic_model
                  end
 
   matching_models = bedrock_models.select do |bedrock_model|
-    model_without_prefix = bedrock_model.sub(/^(?:us\.)?anthropic\./, '')
-    model_without_prefix.start_with?(base_pattern)
+    model_without_prefix = bedrock_model.sub(/^(?:(?:[a-z]{2}|global)\.)?anthropic\./, '')
+    model_without_prefix.match?(/\A#{Regexp.escape(base_pattern)}(?:-v\d+|:\d+k|$)/)
   end
 
   return nil if matching_models.empty?

data/lib/tasks/release.rake

@@ -1,5 +1,48 @@
 # frozen_string_literal: true
 
+require 'time'
+require 'yaml'
+
+# Shared helpers for release-related Rake tasks.
+module ReleaseTasks
+  module_function
+
+  def cassette_recorded_at_times(cassette)
+    data = YAML.safe_load_file(cassette, aliases: true)
+
+    Array(data['http_interactions']).filter_map do |interaction|
+      Time.parse(interaction['recorded_at'])
+    rescue ArgumentError, TypeError
+      nil
+    end
+  rescue Psych::Exception => e
+    abort "Could not parse VCR cassette #{cassette}: #{e.message}"
+  end
+
+  def cassette_recorded_at(cassette)
+    recorded_at_times = cassette_recorded_at_times(cassette)
+
+    abort "No recorded_at timestamps found in VCR cassette #{cassette}" if recorded_at_times.empty?
+
+    recorded_at_times.min
+  end
+
+  def find_stale_cassettes(cassette_dir, max_age_days)
+    Dir.glob("#{cassette_dir}/**/*.yml").filter_map do |cassette|
+      recorded_at = cassette_recorded_at(cassette)
+      age_days = (Time.now - recorded_at) / 86_400
+
+      next unless age_days > max_age_days
+
+      {
+        path: cassette,
+        file: File.basename(cassette),
+        age: age_days.round(1)
+      }
+    end
+  end
+end
+
 namespace :release do # rubocop:disable Metrics/BlockLength
   desc 'Prepare for release'
   task :prepare do
@@ -13,18 +56,15 @@ namespace :release do # rubocop:disable Metrics/BlockLength
     max_age_days = 1
     cassette_dir = 'spec/fixtures/vcr_cassettes'
 
-    stale_count = 0
-    Dir.glob("#{cassette_dir}/**/*.yml").each do |cassette|
-      age_days = (Time.now - File.mtime(cassette)) / 86_400
-      next unless age_days > max_age_days
+    stale_cassettes = ReleaseTasks.find_stale_cassettes(cassette_dir, max_age_days)
 
-      puts "Removing stale cassette: #{File.basename(cassette)} (#{age_days.round(1)} days old)"
-      File.delete(cassette)
-      stale_count += 1
+    stale_cassettes.each do |cassette|
+      puts "Removing stale cassette: #{cassette[:file]} (#{cassette[:age]} days old)"
+      File.delete(cassette[:path])
     end
 
-    if stale_count.positive?
-      puts "\n🗑️  Removed #{stale_count} stale cassettes"
+    if stale_cassettes.any?
+      puts "\n🗑️  Removed #{stale_cassettes.size} stale cassettes"
       puts '🔄 Re-recording cassettes...'
       run_test_queue_rspec || exit(1)
       puts '✅ Cassettes refreshed!'
@@ -37,25 +77,12 @@ namespace :release do # rubocop:disable Metrics/BlockLength
   task :verify_cassettes do
     max_age_days = 1
     cassette_dir = 'spec/fixtures/vcr_cassettes'
-    stale_cassettes = []
-
-    Dir.glob("#{cassette_dir}/**/*.yml").each do |cassette|
-      age_days = (Time.now - File.mtime(cassette)) / 86_400
-
-      next unless age_days > max_age_days
-
-      stale_cassettes << {
-        file: File.basename(cassette),
-        age: age_days.round(1)
-      }
-    end
+    stale_cassettes = ReleaseTasks.find_stale_cassettes(cassette_dir, max_age_days)
 
     if stale_cassettes.any?
       puts "\n❌ Found stale cassettes (older than #{max_age_days} days):"
-      stale_files = []
       stale_cassettes.each do |c|
         puts "   - #{c[:file]} (#{c[:age]} days old)"
-        stale_files << File.join(cassette_dir, '**', c[:file])
       end
 
       puts "\nRun locally: bundle exec rake release:refresh_stale_cassettes"

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: ruby_llm
 version: !ruby/object:Gem::Version
-  version: 1.14.1
+  version: 1.16.0
 platform: ruby
 authors:
 - Carmine Paolino
@@ -135,13 +135,13 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '2'
-description: One beautiful Ruby API for GPT, Claude, Gemini, and more. Easily build
-  chatbots, AI agents, RAG applications, and content generators. Features chat (text,
-  images, audio, PDFs), image generation, embeddings, tools (function calling), structured
-  output, Rails integration, and streaming. Works with OpenAI, Anthropic, Google Gemini,
-  AWS Bedrock, DeepSeek, Mistral, Ollama (local models), OpenRouter, Perplexity, GPUStack,
-  and any OpenAI-compatible API. Minimal dependencies - just Faraday, Zeitwerk, and
-  Marcel.
+description: A single, beautiful Ruby framework for all major AI providers. Easily
+  build chatbots, AI agents, RAG applications, content generators, and every AI workflow
+  you can think of. Features chat (text, images, audio, PDFs), image generation, embeddings,
+  tools (function calling), structured output, Rails integration, and streaming. Works
+  with OpenAI, Anthropic, Google Gemini, AWS Bedrock, DeepSeek, Mistral, Ollama (local
+  models), OpenRouter, Perplexity, GPUStack, and any OpenAI-compatible API. Minimal
+  dependencies - just Faraday, Zeitwerk, and Marcel.
 email:
 - carmine@paolino.me
 executables: []
@@ -240,9 +240,13 @@ files:
 - lib/ruby_llm/connection.rb
 - lib/ruby_llm/content.rb
 - lib/ruby_llm/context.rb
+- lib/ruby_llm/cost.rb
+- lib/ruby_llm/deprecator.rb
 - lib/ruby_llm/embedding.rb
 - lib/ruby_llm/error.rb
+- lib/ruby_llm/error_middleware.rb
 - lib/ruby_llm/image.rb
+- lib/ruby_llm/instrumentation.rb
 - lib/ruby_llm/message.rb
 - lib/ruby_llm/mime_type.rb
 - lib/ruby_llm/model.rb
@@ -251,6 +255,7 @@ files:
 - lib/ruby_llm/model/pricing.rb
 - lib/ruby_llm/model/pricing_category.rb
 - lib/ruby_llm/model/pricing_tier.rb
+- lib/ruby_llm/model_registry.rb
 - lib/ruby_llm/models.json
 - lib/ruby_llm/models.rb
 - lib/ruby_llm/models_schema.json
@@ -298,6 +303,7 @@ files:
 - lib/ruby_llm/providers/mistral/capabilities.rb
 - lib/ruby_llm/providers/mistral/chat.rb
 - lib/ruby_llm/providers/mistral/embeddings.rb
+- lib/ruby_llm/providers/mistral/media.rb
 - lib/ruby_llm/providers/mistral/models.rb
 - lib/ruby_llm/providers/ollama.rb
 - lib/ruby_llm/providers/ollama/capabilities.rb
@@ -324,6 +330,7 @@ files:
 - lib/ruby_llm/providers/perplexity.rb
 - lib/ruby_llm/providers/perplexity/capabilities.rb
 - lib/ruby_llm/providers/perplexity/chat.rb
+- lib/ruby_llm/providers/perplexity/media.rb
 - lib/ruby_llm/providers/perplexity/models.rb
 - lib/ruby_llm/providers/vertexai.rb
 - lib/ruby_llm/providers/vertexai/chat.rb
@@ -341,6 +348,7 @@ files:
 - lib/ruby_llm/tokens.rb
 - lib/ruby_llm/tool.rb
 - lib/ruby_llm/tool_call.rb
+- lib/ruby_llm/tool_concurrency.rb
 - lib/ruby_llm/transcription.rb
 - lib/ruby_llm/utils.rb
 - lib/ruby_llm/version.rb
@@ -354,14 +362,28 @@ licenses:
 metadata:
   homepage_uri: https://rubyllm.com
   source_code_uri: https://github.com/crmne/ruby_llm
-  changelog_uri: https://github.com/crmne/ruby_llm/commits/main
+  changelog_uri: https://github.com/crmne/ruby_llm/releases
   documentation_uri: https://rubyllm.com
   bug_tracker_uri: https://github.com/crmne/ruby_llm/issues
   funding_uri: https://github.com/sponsors/crmne
   rubygems_mfa_required: 'true'
 post_install_message: |
-  Upgrading from RubyLLM < 1.14.x? Check the upgrade guide for new features and migration instructions
-  --> https://rubyllm.com/upgrading/
+  RubyLLM 1.15 upgrade note:
+
+    Token accounting is now normalized across providers. `input_tokens` means
+    standard input tokens; prompt cache reads and writes are exposed separately
+    as `cache_read_tokens` and `cache_write_tokens`.
+
+    Need request-side input activity?
+      input_tokens + cache_read_tokens + cache_write_tokens
+
+    New cost helpers:
+      response.cost.total
+      chat.cost.total
+      agent.cost.total
+
+    Upgrading from RubyLLM < 1.15? Read the full upgrade guide:
+      https://rubyllm.com/upgrading/
 rdoc_options: []
 require_paths:
 - lib
@@ -376,7 +398,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 4.0.6
+rubygems_version: 4.0.10
 specification_version: 4
-summary: One beautiful Ruby API for GPT, Claude, Gemini, and more.
+summary: A single, beautiful Ruby framework for all major AI providers.
 test_files: []