Model Metadata
- name
- Kimi K3
- description
- Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/gif, image/webp, image/bmp, image/heic, image/heif, video/mp4, video/mpeg, video/mov, video/avi, video/x-flv, video/mpg, video/webm, video/wmv, video/3gpp; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; temperature: No; top_p: No; top_k: No
- limit
- context: 1,048,576; output: 131,072
Provider Variants 23
Alibaba kimi-k3
Context 1,048,576 · Output 1,048,576
- name
- Kimi K3
- description
- Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/gif, image/webp, image/bmp, image/heic, image/heif; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; default: type: effort; value: max; toggle: Yes; effort: low, high, max; cache: implicit: min_tokens: 1,024; structured_output: Yes; temperature: No; top_p: No; top_k: No; system_updates: No; effort_updates: No; stream_usage: Yes
- limit
- context: 1,048,576; output: 1,048,576
- id
- kimi-k3
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- api
- chat-completions: base_url: https://dashscope-intl.aliyuncs.com/compatible-mode/v1; sdk: aisdk: @ai-sdk/openai-compatible; opencodeai: @opencode/ai/providers/alibaba/chat; compatibility: reasoning_field: reasoning_content; capabilities: system_updates: Yes; tools: choice: auto, none, required, tool; updates: No; responses: base_url: https://dashscope-intl.aliyuncs.com/compatible-mode/v1; sdk: aisdk: @ai-sdk/openai; opencodeai: @opencode/ai/providers/alibaba/responses; capabilities: tools: choice: auto, none, tool; updates: No; messages: base_url: https://dashscope-intl.aliyuncs.com/apps/anthropic/v1; sdk: aisdk: @ai-sdk/anthropic; opencodeai: @opencode/ai/providers/alibaba/messages; capabilities: tools: choice: auto, none, required, tool; updates: No
amazon-bedrock global.moonshotai.kimi-k3
Context 1,048,576 · Output 128,000
- name
- Kimi K3 (Global)
- description
- Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/gif, image/webp; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; cache: implicit: Yes; structured_output: Yes; temperature: No; top_p: No; top_k: No; system_updates: No; effort_updates: No; service_tier: flex, priority
- limit
- context: 1,048,576; output: 128,000
- id
- global.moonshotai.kimi-k3
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- api
- converse: base_url: https://bedrock-runtime.${AWS_REGION}.amazonaws.com; sdk: aisdk: @ai-sdk/amazon-bedrock; opencodeai: @opencode/ai/providers/amazon-bedrock; capabilities: tools: strict: Yes; updates: No
amazon-bedrock us.moonshotai.kimi-k3
Context 1,048,576 · Output 128,000
- name
- Kimi K3 (US)
- description
- Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/gif, image/webp; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; cache: implicit: Yes; structured_output: Yes; temperature: No; top_p: No; top_k: No; system_updates: No; effort_updates: No; service_tier: flex, priority
- limit
- context: 1,048,576; output: 128,000
- id
- us.moonshotai.kimi-k3
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- api
- converse: base_url: https://bedrock-runtime.${AWS_REGION}.amazonaws.com; sdk: aisdk: @ai-sdk/amazon-bedrock; opencodeai: @opencode/ai/providers/amazon-bedrock; capabilities: tools: strict: Yes; updates: No
arcee moonshotai/kimi-k3
Context 1,048,576 · Output 131,072
- name
- Kimi K3
- description
- Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; effort: low, high, max; structured_output: Yes; temperature: Yes; top_p: No; top_k: No; system_updates: No; effort_updates: No
- limit
- context: 1,048,576; output: 131,072
- id
- moonshotai/kimi-k3
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- status
- beta
- api
- chat-completions: base_url: https://api.arcee.ai/api/v1; sdk: aisdk: @ai-sdk/openai-compatible; opencodeai: @opencode/ai/providers/openai-compatible; compatibility: reasoning_field: reasoning_content; max_tokens_field: max_tokens; capabilities: tools: updates: No
Baseten moonshotai/Kimi-K3
Context 1,048,576 · Output 262,144
- name
- Kimi K3
- description
- Kimi multimodal agent model for visual understanding, coding, and planning
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/png, image/jpeg, image/gif, image/webp, image/bmp, image/tiff; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; effort: none, low, high, max; cache: implicit: Yes; structured_output: Yes; temperature: Yes; top_p: No; top_k: No; system_updates: Yes; effort_updates: No
- limit
- context: 1,048,576; output: 262,144
- id
- moonshotai/Kimi-K3
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- api
- chat-completions: base_url: https://inference.baseten.co/v1; sdk: aisdk: @ai-sdk/openai-compatible; opencodeai: @opencode/ai/providers/baseten; compatibility: reasoning_field: reasoning_content; capabilities: tools: updates: No; messages: base_url: https://inference.baseten.co/v1; sdk: aisdk: @ai-sdk/anthropic; opencodeai: @opencode/ai/providers/anthropic-compatible; capabilities: tools: updates: No
cloudflare-ai-gateway moonshotai/kimi-k3
Context 1,048,576 · Output 131,072
- name
- Kimi K3
- description
- Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/gif, image/webp, image/bmp, image/heic, image/heif, video/mp4, video/mpeg, video/mov, video/avi, video/x-flv, video/mpg, video/webm, video/wmv, video/3gpp; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; effort: none, low, high, max; cache: implicit: Yes; structured_output: Yes; temperature: No; top_p: No; top_k: No; system_updates: Yes; effort_updates: No
- limit
- context: 1,048,576; output: 131,072
- id
- moonshotai/kimi-k3
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- api
- chat-completions: base_url: https://api.cloudflare.com/client/v4/accounts/${CLOUDFLARE_ACCOUNT_ID}/ai/v1; sdk: aisdk: ai-gateway-provider; opencodeai: @opencode/ai/providers/cloudflare-ai-gateway; capabilities: tools: updates: Yes
deepinfra moonshotai/Kimi-K3
Context 1,048,576 · Output 131,072
- name
- Kimi K3
- description
- Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/webp; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; effort: low, high, max; cache: implicit: Yes; explicit: ttl: 5m, 1h; min_tokens: 1,536; structured_output: Yes; temperature: No; top_p: No; top_k: No; system_updates: No; effort_updates: No; prompt_cache_key: Yes
- limit
- context: 1,048,576; output: 131,072
- id
- moonshotai/Kimi-K3
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- api
- chat-completions: base_url: https://api.deepinfra.com/v1/openai; sdk: aisdk: @ai-sdk/deepinfra; opencodeai: @opencode/ai/providers/deepinfra; capabilities: tools: updates: No
Fireworks accounts/fireworks/models/kimi-k3
Context 1,048,576 · Output 131,072
- name
- Kimi K3
- description
- Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/gif, image/webp, image/bmp; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; toggle: Yes; effort: low, medium, high, max; budget: min: 1,024; cache: implicit: Yes; structured_output: Yes; temperature: No; top_p: No; top_k: No; system_updates: Yes; effort_updates: No
- limit
- context: 1,048,576; output: 131,072
- id
- accounts/fireworks/models/kimi-k3
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- api
- chat-completions: base_url: https://api.fireworks.ai/inference/v1; sdk: aisdk: @ai-sdk/openai-compatible; opencodeai: @opencode/ai/providers/fireworks; compatibility: reasoning_field: reasoning_content; capabilities: tools: updates: No; responses: base_url: https://api.fireworks.ai/inference/v1; sdk: aisdk: @ai-sdk/openai; opencodeai: @opencode/ai/providers/openai-compatible/responses; capabilities: tools: updates: No; messages: base_url: https://api.fireworks.ai/inference/v1; sdk: aisdk: @ai-sdk/anthropic; opencodeai: @opencode/ai/providers/anthropic-compatible; capabilities: tools: updates: No
- experimental
- modes: priority: body: service_tier: priority
Fireworks accounts/fireworks/routers/kimi-fast-latest
Context 1,048,576 · Output 131,072
- name
- Kimi Fast Latest
- description
- Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/gif, image/webp, image/bmp; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; toggle: Yes; effort: low, medium, high, max; budget: min: 1,024; cache: implicit: Yes; structured_output: Yes; temperature: No; top_p: No; top_k: No; system_updates: Yes; effort_updates: No
- limit
- context: 1,048,576; output: 131,072
- id
- accounts/fireworks/routers/kimi-fast-latest
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- api
- chat-completions: base_url: https://api.fireworks.ai/inference/v1; sdk: aisdk: @ai-sdk/openai-compatible; opencodeai: @opencode/ai/providers/fireworks; compatibility: reasoning_field: reasoning_content; capabilities: tools: updates: No; responses: base_url: https://api.fireworks.ai/inference/v1; sdk: aisdk: @ai-sdk/openai; opencodeai: @opencode/ai/providers/openai-compatible/responses; capabilities: tools: updates: No; messages: base_url: https://api.fireworks.ai/inference/v1; sdk: aisdk: @ai-sdk/anthropic; opencodeai: @opencode/ai/providers/anthropic-compatible; capabilities: tools: updates: No
Fireworks accounts/fireworks/routers/kimi-k3-fast
Context 1,048,576 · Output 131,072
- name
- Kimi K3 Fast
- description
- Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/gif, image/webp, image/bmp; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; toggle: Yes; effort: low, medium, high, max; budget: min: 1,024; cache: implicit: Yes; structured_output: Yes; temperature: No; top_p: No; top_k: No; system_updates: Yes; effort_updates: No
- limit
- context: 1,048,576; output: 131,072
- id
- accounts/fireworks/routers/kimi-k3-fast
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- api
- chat-completions: base_url: https://api.fireworks.ai/inference/v1; sdk: aisdk: @ai-sdk/openai-compatible; opencodeai: @opencode/ai/providers/fireworks; compatibility: reasoning_field: reasoning_content; capabilities: tools: updates: No; responses: base_url: https://api.fireworks.ai/inference/v1; sdk: aisdk: @ai-sdk/openai; opencodeai: @opencode/ai/providers/openai-compatible/responses; capabilities: tools: updates: No; messages: base_url: https://api.fireworks.ai/inference/v1; sdk: aisdk: @ai-sdk/anthropic; opencodeai: @opencode/ai/providers/anthropic-compatible; capabilities: tools: updates: No
Fireworks accounts/fireworks/routers/kimi-k3-spot
Context 1,048,576 · Output 131,072
- name
- Kimi K3 Spot
- description
- Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/gif, image/webp, image/bmp; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; toggle: Yes; effort: low, medium, high, max; budget: min: 1,024; cache: implicit: Yes; structured_output: Yes; temperature: No; top_p: No; top_k: No; system_updates: Yes; effort_updates: No
- limit
- context: 1,048,576; output: 131,072
- id
- accounts/fireworks/routers/kimi-k3-spot
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- api
- chat-completions: base_url: https://api.fireworks.ai/inference/v1; sdk: aisdk: @ai-sdk/openai-compatible; opencodeai: @opencode/ai/providers/fireworks; compatibility: reasoning_field: reasoning_content; capabilities: tools: updates: No; responses: base_url: https://api.fireworks.ai/inference/v1; sdk: aisdk: @ai-sdk/openai; opencodeai: @opencode/ai/providers/openai-compatible/responses; capabilities: tools: updates: No; messages: base_url: https://api.fireworks.ai/inference/v1; sdk: aisdk: @ai-sdk/anthropic; opencodeai: @opencode/ai/providers/anthropic-compatible; capabilities: tools: updates: No
Fireworks accounts/fireworks/routers/kimi-latest
Context 1,048,576 · Output 131,072
- name
- Kimi Latest
- description
- Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/gif, image/webp, image/bmp; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; toggle: Yes; effort: low, medium, high, max; budget: min: 1,024; cache: implicit: Yes; structured_output: Yes; temperature: No; top_p: No; top_k: No; system_updates: Yes; effort_updates: No
- limit
- context: 1,048,576; output: 131,072
- id
- accounts/fireworks/routers/kimi-latest
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- api
- chat-completions: base_url: https://api.fireworks.ai/inference/v1; sdk: aisdk: @ai-sdk/openai-compatible; opencodeai: @opencode/ai/providers/fireworks; compatibility: reasoning_field: reasoning_content; capabilities: tools: updates: No; responses: base_url: https://api.fireworks.ai/inference/v1; sdk: aisdk: @ai-sdk/openai; opencodeai: @opencode/ai/providers/openai-compatible/responses; capabilities: tools: updates: No; messages: base_url: https://api.fireworks.ai/inference/v1; sdk: aisdk: @ai-sdk/anthropic; opencodeai: @opencode/ai/providers/anthropic-compatible; capabilities: tools: updates: No
- experimental
- modes: priority: body: service_tier: priority
github-copilot kimi-k3
Context 1,048,576 · Output 131,072
- name
- Kimi K3
- description
- Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/webp, image/gif; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; effort: low, high, max; structured_output: Yes; temperature: No; top_p: No; top_k: No; system_updates: No; effort_updates: No
- limit
- context: 1,048,576; output: 131,072
- id
- kimi-k3
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- api
- chat-completions: base_url: https://api.githubcopilot.com; sdk: aisdk: @ai-sdk/openai-compatible; opencodeai: @opencode/ai/providers/openai-compatible; capabilities: tools: updates: No
huggingface moonshotai/Kimi-K3
Context 1,000,000 · Output 131,072
- name
- Kimi K3
- description
- Kimi multimodal agent model for visual understanding, coding, and planning
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/gif, image/webp, image/bmp, image/heic, image/heif; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; effort: low, high, max; structured_output: Yes; temperature: No; top_p: No; top_k: No; system_updates: No; effort_updates: No
- limit
- context: 1,000,000; output: 131,072
- id
- moonshotai/Kimi-K3
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- api
- chat-completions: base_url: https://router.huggingface.co/v1; sdk: aisdk: @ai-sdk/openai-compatible; opencodeai: @opencode/ai/providers/openai-compatible; compatibility: max_tokens_field: max_tokens; capabilities: tools: updates: No; responses: base_url: https://router.huggingface.co/v1; sdk: aisdk: @ai-sdk/openai; opencodeai: @opencode/ai/providers/openai-compatible/responses; capabilities: tools: updates: No
Moonshot AI kimi-k3
Context 1,048,576 · Output 1,048,576
- name
- Kimi K3
- description
- Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/gif, image/webp, image/bmp, image/heic, image/heif, video/mp4, video/mpeg, video/mov, video/avi, video/x-flv, video/mpg, video/webm, video/wmv, video/3gpp; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; default: type: effort; value: max; effort: low, high, max; cache: implicit: ttl: 5m, 1h; structured_output: Yes; temperature: No; top_p: No; top_k: No; system_updates: No; effort_updates: No; stream_usage: Yes; prompt_cache_key: Yes
- limit
- context: 1,048,576; output: 1,048,576
- id
- kimi-k3
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- api
- chat-completions: base_url: https://api.moonshot.ai/v1; sdk: aisdk: @ai-sdk/openai-compatible; opencodeai: @opencode/ai/providers/moonshot/chat; compatibility: reasoning_field: reasoning_content; capabilities: system_updates: Yes; tools: strict: Yes; choice: auto, none, required; updates: Yes; responses: base_url: https://api.moonshot.ai/v1; sdk: aisdk: @ai-sdk/openai; opencodeai: @opencode/ai/providers/moonshot/responses; capabilities: tools: strict: Yes; choice: auto; updates: No; messages: base_url: https://api.moonshot.ai/anthropic/v1; sdk: aisdk: @ai-sdk/anthropic; opencodeai: @opencode/ai/providers/moonshot/messages; capabilities: tools: strict: Yes; choice: auto, none, required; updates: No
NVIDIA moonshotai/kimi-k3
Context 1,048,576 · Output 131,072
- name
- Kimi K3
- description
- Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/gif, image/webp, image/bmp, image/heic, image/heif, video/mp4, video/mpeg, video/mov, video/avi, video/x-flv, video/mpg, video/webm, video/wmv, video/3gpp; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; toggle: Yes; effort: low, high, max; structured_output: Yes; temperature: No; top_p: No; top_k: No; system_updates: No; effort_updates: No
- limit
- context: 1,048,576; output: 131,072
- id
- moonshotai/kimi-k3
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- status
- deprecated
- api
- chat-completions: base_url: https://integrate.api.nvidia.com/v1; sdk: aisdk: @ai-sdk/openai-compatible; opencodeai: @opencode/ai/providers/openai-compatible; compatibility: reasoning_field: reasoning_content; capabilities: tools: updates: No
ollama-cloud kimi-k3
Context 1,051,648 · Output 131,072
Amp's provider snapshot overrides input/output limits and image support for this serving path. Other fields use models.dev.
- name
- kimi-k3
- description
- Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/gif, image/webp, image/bmp, image/heic, image/heif; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; toggle: Yes; effort: low, high, max; cache: implicit: Yes; structured_output: No; temperature: No; top_p: No; top_k: No; system_updates: No; effort_updates: No
- limit
- context: 1,051,648; output: 131,072; input: 920,576
- id
- kimi-k3
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- api
- chat-completions: base_url: https://ollama.com/v1; sdk: aisdk: @ai-sdk/openai-compatible; opencodeai: @opencode/ai/providers/openai-compatible; compatibility: max_tokens_field: max_tokens; capabilities: tools: updates: No; responses: base_url: https://ollama.com/v1; sdk: aisdk: @ai-sdk/openai; opencodeai: @opencode/ai/providers/openai-compatible/responses; capabilities: tools: updates: No; messages: base_url: https://ollama.com/v1; sdk: aisdk: @ai-sdk/anthropic; opencodeai: @opencode/ai/providers/anthropic-compatible; capabilities: tools: updates: No
opencode kimi-k3
Context 1,048,576 · Output 131,072
- name
- Kimi K3
- description
- Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/gif, image/webp, image/bmp, image/heic, image/heif, video/mp4, video/mpeg, video/mov, video/avi, video/x-flv, video/mpg, video/webm, video/wmv, video/3gpp; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; toggle: Yes; effort: low, high, max; cache: implicit: Yes; structured_output: Yes; temperature: No; top_p: No; top_k: No; system_updates: Yes; effort_updates: No
- limit
- context: 1,048,576; output: 131,072
- id
- kimi-k3
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- api
- chat-completions: base_url: https://opencode.ai/inference/openai/v1; sdk: aisdk: @ai-sdk/openai-compatible; opencodeai: @opencode/ai/providers/openai-compatible; compatibility: reasoning_field: reasoning_content; capabilities: tools: strict: Yes; choice: auto, none, required; updates: Yes
opencode-go kimi-k3
Context 1,051,648 · Output 131,072
Amp's provider snapshot overrides input/output limits and image support for this serving path. Other fields use models.dev.
- name
- Kimi K3
- description
- Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/gif, image/webp, image/bmp, image/heic, image/heif, video/mp4, video/mpeg, video/mov, video/avi, video/x-flv, video/mpg, video/webm, video/wmv, video/3gpp; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; effort: max; structured_output: Yes; temperature: No; top_p: No; top_k: No; system_updates: Yes; effort_updates: No
- limit
- context: 1,051,648; output: 131,072; input: 920,576
- id
- kimi-k3
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- api
- chat-completions: base_url: https://opencode.ai/inference/go/openai/v1; sdk: aisdk: @ai-sdk/openai-compatible; opencodeai: @opencode/ai/providers/openai-compatible; compatibility: reasoning_field: reasoning_content; capabilities: tools: strict: Yes; choice: auto, none, required; updates: Yes
OpenRouter moonshotai/kimi-k3
Context 1,048,576 · Output 128,000
Amp's provider snapshot overrides input/output limits and image support for this serving path. Other fields use models.dev.
- name
- Kimi K3
- description
- Kimi multimodal agent model for visual understanding, coding, and planning
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/gif, image/webp, image/bmp, image/heic, image/heif, video/mp4, video/mpeg, video/mov, video/avi, video/x-flv, video/mpg, video/webm, video/wmv, video/3gpp; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; effort: none, low, high, max; structured_output: Yes; temperature: Yes; top_p: No; top_k: No; system_updates: No; effort_updates: No
- limit
- context: 1,048,576; output: 128,000; input: 920,576
- id
- moonshotai/kimi-k3
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- api
- chat-completions: base_url: https://openrouter.ai/api/v1; sdk: aisdk: @openrouter/ai-sdk-provider; opencodeai: @opencode/ai/providers/openrouter; capabilities: tools: choice: auto, none, required, tool; updates: No; responses: base_url: https://openrouter.ai/api/v1; sdk: aisdk: @ai-sdk/openai; opencodeai: @opencode/ai/providers/openai-compatible/responses; capabilities: tools: choice: auto, none, required, tool; updates: No; messages: base_url: https://openrouter.ai/api/v1; sdk: aisdk: @ai-sdk/anthropic; opencodeai: @opencode/ai/providers/anthropic-compatible; capabilities: tools: choice: auto, none, required, tool; updates: No
togetherai moonshotai/Kimi-K3
Context 1,048,576 · Output 131,072
- name
- Kimi K3
- description
- Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/gif, image/webp; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; effort: none, low, high, max; cache: implicit: Yes; structured_output: Yes; temperature: No; top_p: No; top_k: No; system_updates: Yes; effort_updates: No; prompt_cache_key: Yes
- limit
- context: 1,048,576; output: 131,072
- id
- moonshotai/Kimi-K3
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- api
- chat-completions: base_url: https://api.together.xyz/v1; sdk: aisdk: @ai-sdk/togetherai; opencodeai: @opencode/ai/providers/togetherai; compatibility: reasoning_field: reasoning_content; capabilities: tools: updates: No
vercel moonshotai/kimi-k3
Context 1,003,072 · Output 131,072
Amp's provider snapshot overrides input/output limits and image support for this serving path. Other fields use models.dev.
- name
- Kimi K3
- description
- Kimi multimodal agent model for visual understanding, coding, and planning
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/gif, image/webp, image/bmp, image/heic, image/heif, application/pdf; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; effort: low, high, max; cache: implicit: Yes; explicit: Yes; structured_output: Yes; temperature: No; top_p: No; top_k: No; system_updates: No; effort_updates: No
- limit
- context: 1,003,072; output: 131,072; input: 872,000
- id
- moonshotai/kimi-k3
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- api
- chat-completions: base_url: https://ai-gateway.vercel.sh/v1; sdk: aisdk: @ai-sdk/gateway; opencodeai: @opencode/ai/providers/vercel-ai-gateway; capabilities: system_updates: Yes; tools: choice: auto, none, required, tool; updates: No; responses: base_url: https://ai-gateway.vercel.sh/v1; sdk: aisdk: @ai-sdk/open-responses; opencodeai: @opencode/ai/providers/vercel-ai-gateway/responses; capabilities: tools: choice: auto, none, required, tool; updates: No; messages: base_url: https://ai-gateway.vercel.sh/v1; sdk: aisdk: @ai-sdk/anthropic; opencodeai: @opencode/ai/providers/vercel-ai-gateway/messages; capabilities: tools: choice: auto, none, required, tool; updates: No
vercel moonshotai/kimi-k3-fast
Context 1,003,072 · Output 131,072
Amp's provider snapshot overrides input/output limits and image support for this serving path. Other fields use models.dev.
- name
- Kimi K3 Fast
- description
- Kimi multimodal agent model for visual understanding, coding, and planning
- family
- kimi-k3
- open weights
- Yes
- release date
- 2026-07-16
- modalities
- input: text, image/jpeg, image/png, image/gif, image/webp, image/bmp, image/heic, image/heif, application/pdf; output: text
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; effort: none, low, medium, high, max; cache: implicit: Yes; structured_output: Yes; temperature: No; top_p: No; top_k: No; system_updates: No; effort_updates: No
- limit
- context: 1,003,072; output: 131,072; input: 872,000
- id
- moonshotai/kimi-k3-fast
- canonical id
- moonshotai/kimi-k3
- type
- chat
- last updated
- 2026-07-16
- api
- chat-completions: base_url: https://ai-gateway.vercel.sh/v1; sdk: aisdk: @ai-sdk/gateway; opencodeai: @opencode/ai/providers/vercel-ai-gateway; capabilities: system_updates: Yes; tools: choice: auto, none, required, tool; updates: No; responses: base_url: https://ai-gateway.vercel.sh/v1; sdk: aisdk: @ai-sdk/open-responses; opencodeai: @opencode/ai/providers/vercel-ai-gateway/responses; capabilities: tools: choice: auto, none, required, tool; updates: No; messages: base_url: https://ai-gateway.vercel.sh/v1; sdk: aisdk: @ai-sdk/anthropic; opencodeai: @opencode/ai/providers/vercel-ai-gateway/messages; capabilities: tools: choice: auto, none, required, tool; updates: No