Skip to content

[Feat] Add Kimi K3 model to Moonshot and OpenCode Go providers #996

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
navedmerchant merged 3 commits into from
Jul 24, 2026
Merged

[Feat] Add Kimi K3 model to Moonshot and OpenCode Go providers #996

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
f685cc2
feat(types): add Kimi K3 to moonshot and opencode-go providers
roomote Jul 24, 2026
2517179
test(types): add moonshot registry suite and explicit kimi-k3 reasoni…
roomote Jul 24, 2026
e782967
test(types): widen moonshot registry entries to ModelInfo in tests
roomote Jul 24, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions packages/types/src/__tests__/lite-llm.test.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,6 +12,7 @@ describe("LiteLLM preserveReasoning model detection", () => {
})

it("matches provider-prefixed routed model names by their final segment", () => {
expect(isLiteLLMPreserveReasoningModel("kimi-k3")).toBe(true)
expect(isLiteLLMPreserveReasoningModel("deepseek/deepseek-reasoner")).toBe(true)
expect(isLiteLLMPreserveReasoningModel("bedrock/moonshot.kimi-k2-thinking")).toBe(true)
expect(isLiteLLMPreserveReasoningModel("fireworks_ai/accounts/fireworks/models/kimi-k2p7-code")).toBe(true)
Expand Down
1 change: 1 addition & 0 deletions packages/types/src/__tests__/opencode-go.test.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,6 +21,7 @@ describe("opencode-go registry", () => {
"glm-5",
"glm-5.1",
"glm-5.2",
"kimi-k3",
"kimi-k2.5",
Comment thread
zoomote[bot] marked this conversation as resolved.
"kimi-k2.6",
"mimo-v2.5",
Expand Down
3 changes: 3 additions & 0 deletions packages/types/src/providers/lite-llm.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -48,6 +48,9 @@ export const LITELLM_PRESERVE_REASONING_MODEL_IDS = [
"moonshot.kimi-k2-thinking",
"kimi-k2p7-code",

// moonshot.ts, opencode-go.ts
"kimi-k3",

// zai.ts
"glm-4.7",
"glm-5",
Expand Down
15 changes: 15 additions & 0 deletions packages/types/src/providers/moonshot.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,6 +6,21 @@ export type MoonshotModelId = keyof typeof moonshotModels
export const moonshotDefaultModelId: MoonshotModelId = "kimi-k2-0905-preview"

export const moonshotModels = {
"kimi-k3": {
Comment thread
zoomote[bot] marked this conversation as resolved.
maxTokens: 131_072, // Default max_completion_tokens (configurable up to 1,048,576)
contextWindow: 1_048_576, // 1M tokens
supportsImages: true, // Native visual understanding (text, image, video)
supportsPromptCache: true, // Automatic context caching
supportsReasoningEffort: ["low", "high", "max"], // Always reasons; default "max"
reasoningEffort: "max",
preserveReasoning: true,
inputPrice: 3.0, // $3.00 per million tokens (cache miss)
outputPrice: 15.0, // $15.00 per million tokens
cacheWritesPrice: 0, // $0 per million tokens (cache miss)
cacheReadsPrice: 0.3, // $0.30 per million tokens (cache hit)
defaultTemperature: 1.0, // temperature is fixed at 1.0
description: `Kimi K3 is Kimi's most capable flagship model with 2.8 trillion parameters, native visual understanding, and a 1M-token context window, designed for long-horizon coding, knowledge work, and deep reasoning. Thinking is always enabled with configurable reasoning effort (low/high/max, default max).`,
},
"kimi-k2-0711-preview": {
maxTokens: 32_000,
contextWindow: 131_072,
Expand Down
17 changes: 17 additions & 0 deletions packages/types/src/providers/opencode-go.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -104,6 +104,23 @@ export const opencodeGoModels: Record<string, ModelInfo> = {
},

// --- Moonshot Kimi ---
"kimi-k3": {
maxTokens: 131_072, // Default max_completion_tokens (configurable up to 1,048,576)
contextWindow: 1_048_576,
supportsImages: false,
supportsPromptCache: true,
supportsMaxTokens: true,
supportsReasoningEffort: ["low", "high", "max"], // Always reasons; default "max"
reasoningEffort: "max",
preserveReasoning: true,
defaultTemperature: 1.0,
// Go pricing matches Moonshot direct ($3 in / $0.30 cache / $15 out per 1M tokens).
inputPrice: 3.0,
outputPrice: 15.0,
cacheReadsPrice: 0.3,
description:
"Kimi K3 is Moonshot AI's flagship model with 2.8 trillion parameters, a 1M context window, and always-on reasoning with configurable effort (low/high/max). Available via the Opencode Go plan.",
},
"kimi-k2.5": {
maxTokens: 16_384,
contextWindow: 262_144,
Expand Down
Loading