Scope

• Base ref: origin/main
• Source ref: opencode/dev
• Theme id: opencode-provider
• Commits mapped: 65
• Date span: 2026-01-21 to 2026-02-13

Theme Summary

Model/provider integration logic and provider-specific behavior.

Mapping Rule

Commits are assigned to one issue by dominant touched-path prefix (most matched files in commit; ties resolved by longest prefix then fixed theme order).

Top Path Hotspots

• packages/opencode/src/provider/transform.ts (34)
• packages/opencode/src/provider/provider.ts (20)
• packages/opencode/test/provider/transform.test.ts (11)
• packages/opencode/src/provider/models.ts (9)
• packages/opencode/src/provider/sdk/copilot/chat/convert-to-openai-compatible-chat-messages.ts (4)
• packages/opencode/test/provider/copilot/convert-to-copilot-messages.test.ts (4)
• packages/opencode/src/session/llm.ts (3)
• packages/opencode/src/provider/sdk/copilot/chat/openai-compatible-chat-language-model.ts (3)
• packages/opencode/test/provider/copilot/copilot-chat-model.test.ts (3)
• bun.lock (3)

Commits

• 0e1a8a18 fix: Claude w/bedrock custom inference profile - caching support (fix: Claude w/bedrock custom inference profile - caching support anomalyco/opencode#9838)
• 6f7a1c69 tweak: adjust textVerbosity and reasoningEffort defaults to better match codex cli
• 8df09abb feat: Make the models.dev domain configurable for offline environments (feat: add OPENCODE_MODELS_URL so the models.dev domain configurable for offline environments anomalyco/opencode#9258)
• 65e267ed feat: Add promptCacheKey for Venice provider (feat: Add promptCacheKey for Venice provider anomalyco/opencode#9915)
• 31094cd5 fix(provider): add thinking presets for Google Vertex Anthropic (fix(provider): add thinking presets for Google Vertex Anthropic anomalyco/opencode#9953)
• 0d9ca0ea sync
• 68bd16df core: fix models snapshot loading to prevent caching issues
• cf7e10c4 fix: add xhigh reasoning effort for GitHub Copilot GPT-5 models (fix: add xhigh reasoning effort for GitHub Copilot GPT-5 models anomalyco/opencode#10092)
• ff9c1864 tests
• 087d7da1 fix(provider): deep merge providerOptions in applyCaching (fix(provider): deep merge providerOptions in applyCaching anomalyco/opencode#10323)
• 0bc4a433 fix(provider): enable thinking for google-vertex-anthropic models (fix(provider): enable thinking for google-vertex-anthropic models anomalyco/opencode#10442)
• 5993a098 fix(core): don't override source in custom provider loaders
• e49306b8 rm log statement
• ac53a372 feat: use anthropic compat messages api for anthropic models through copilot
• 32ce0f4b tweak: add recommended topP/temp for kimi k2.5
• bb63d16f Set temperature for kimi k2.5
• df8b23db Revert "Set temperature for kimi k2.5"
• 0fabdccf fix: ensure that kimi doesnt have fake variants available
• 29ea9fcf fix: ensure variants for copilot models work w/ maxTokens being set
• 301895c7 fix: ensure kimi k2.5 from fireworks ai and kimi for coding providers properly set temperature
• b937fe94 fix(provider): include providerID in SDK cache key (fix(provider): include providerID in SDK cache key anomalyco/opencode#11020)
• 0c32afbc fix(provider): use snake_case for thinking param with OpenAI-compatible APIs (fix(provider): use snake_case for thinking param with OpenAI-compatible APIs anomalyco/opencode#10109)
• 2c36cbb8 refactor(provider): remove google-vertex-anthropic special case in ge… (refactor(provider): remove google-vertex-anthropic special case in ge… anomalyco/opencode#10743)
• 4a56491e fix(provider): exclude chat models from textVerbosity setting (fix(provider): exclude chat models from textVerbosity setting anomalyco/opencode#11363)
• 95bf01a7 fix: ensure the mistral ordering fixes also apply to devstral (fix: ensure the mistral ordering fixes also apply to devstral anomalyco/opencode#11412)
• 252b2c45 chore: generate
• 644f0d4e chore: generate
• 6ecd011e tui: allow specifying custom models file path via OPENCODE_MODELS_PATH
• a19ef17b fix(provider): use process.env directly for runtime env mutations (fix(provider): use process.env directly for runtime env mutations anomalyco/opencode#11482)
• d005e70f core: ensure models configuration is not empty before loading
• d9f18e40 feat(opencode): add copilot specific provider to properly handle copilot reasoning tokens (feat(opencode): add copilot specific provider to properly handle copilot reasoning tokens anomalyco/opencode#8900)
• feca42b0 feat(opencode): add reasoning variants support for SAP AI Core (feat(opencode): add reasoning variants support for SAP AI Core anomalyco/opencode#8753)
• ca5e85d6 fix: prompt caching for opus on bedrock (fix: prompt caching for opus on bedrock anomalyco/opencode#11664)
• d1d74474 fix: ensure switching anthropic models mid convo on copilot works without errors, fix issue with reasoning opaque not being picked up for gemini models (fix: ensure switching anthropic models mid convo on copilot works without errors, fix issue with reasoning opaque not being picked up for gemini models anomalyco/opencode#11569)
• f6948d0f fix: variant logic for anthropic models through openai compat endpoint (fix: variant logic for anthropic models through openai compat endpoint anomalyco/opencode#11665)
• 43354eea fix: convert system message content to string for Copilot provider (fix: convert system message content to string for Copilot provider anomalyco/opencode#11600)
• bd9d7b32 fix: session title generation with OpenAI models. (fix: session title generation with OpenAI models. anomalyco/opencode#11678)
• cf8b033b feat(provider): add User-Agent header for GitLab AI Gateway requests (feat(provider): add User-Agent header for GitLab AI Gateway requests anomalyco/opencode#11818)
• 18585874 chore: generate
• 25bdd77b fix(opencode): use official ai-gateway-provider package for Cloudflare AI Gateway (fix(opencode): use official ai-gateway-provider package for Cloudflare AI Gateway anomalyco/opencode#12014)
• 3741516f fix: handle nested array items for Gemini schema validation (fix: handle nested array items for Gemini schema validation anomalyco/opencode#11952)
• 3adeed8f fix(provider): strip properties/required from non-object types in Gemini schema (fix(provider): strip properties/required from non-object types in Gemini schema anomalyco/opencode#11888)
• 54e14c1a fix: exclude k2p5 from reasoning variants (fix: exclude k2p5 from reasoning variants anomalyco/opencode#11918)
• b7b734f5 fix: ensure mcp tools are sanitized (fix: ensure mcp tools are sanitized anomalyco/opencode#11984)
• 305007aa fix: cloudflare workers ai provider (fix: cloudflare workers ai provider anomalyco/opencode#12157)
• 39a50477 fix: provider headers from config not applied to fetch requests (fix: provider headers from config not applied to fetch requests anomalyco/opencode#11788)
• 5588453c fix: revert change that caused headers to be double merged if provider was authed in multiple places (fix: revert change that caused headers to be double merged if provider was authed in multiple places anomalyco/opencode#12072)
• 8c1f1f13 docs: document the built in agents (docs: document the built in agents anomalyco/opencode#12066)
• b942e0b4 fix: prevent double-prefixing of Bedrock cross-region inference models (fix: prevent double-prefixing of Bedrock cross-region inference models anomalyco/opencode#12056)
• 2f78705f tweak: update transforms for gpt-5.3 (tweak: update transforms for gpt-5.3 anomalyco/opencode#12325)
• d1686661 fix: ensure kimi-for-coding plan has thinking on by default for k2p5 (fix: ensure kimi-for-coding plan has thinking on by default for k2p5 anomalyco/opencode#12147)
• 18749c1f fix(opencode): correct prefix selection for amazon-bedrock provider in getSmallModel (fix(opencode): correct prefix selection for amazon-bedrock provider in getSmallModel anomalyco/opencode#12281)
• 3f7ca049 chore: generate
• a25cd2da feat(opencode): use reasoning summary auto for gpt-5 models that are not chat (feat(opencode): use reasoning summary auto for gpt-5 models that are not chat anomalyco/opencode#12502)
• 99ea1351 tweak: add new ContextOverflowError type (tweak: add new ContextOverflowError type anomalyco/opencode#12777)
• 0cd52f83 fix: enable thinking for all reasoning models on alibaba-cn (DashScope) (fix: enable thinking for all reasoning models on alibaba-cn (DashScope) anomalyco/opencode#12772)
• d40dffb8 fix(provider): remove obsolete copilot model enablement instructions (fix(provider): remove obsolete copilot model enablement instructions anomalyco/opencode#12739)
• d98bd4bd fix: add additional context overflow cases, remove overcorrecting ones (fix: add additional context overflow cases, remove overcorrecting ones anomalyco/opencode#13077)
• 53ec15a5 fix(tui): improve amazon-bedrock check to include container credentials (fix(tui): improve amazon-bedrock check to include container credentials anomalyco/opencode#13037)
• ad208709 support custom api url per model
• 93eee0da fix: look for recent model in fallback in cli (fix: look for recent model in fallback in cli anomalyco/opencode#12582)
• a1155650 core: allow model configurations without npm/api provider details
• bf5a01ed feat(opencode): Venice Add automatic variant generation for Venice models (feat(opencode): Venice Add automatic variant generation for Venice models anomalyco/opencode#12106)
• d475fd61 chore: generate
• 0d90a22f feat: update some ai sdk packages and uuse adaptive reasoning for opus 4.6 on vertex/bedrock/anthropic (feat: update some ai sdk packages and uuse adaptive reasoning for opus 4.6 on vertex/bedrock/anthropic anomalyco/opencode#13439)

Acceptance Checklist

• Validate each commit intent is represented in this theme
• Define parity implementation tasks for Zee based on these commits
• Mark commits that are not applicable to Zee