Model capability probe

MoonshotAI: Kimi K2 Thinking | JSON mode

Whether MoonshotAI: Kimi K2 Thinking supports JSON mode cannot be decided from the name alone. Check model ID moonshotai/kimi-k2-thinking, provider Moonshot AI, catalog params, modality, a lightweight runtime probe, and the error body together.

Model

moonshotai/kimi-k2-thinking

MoonshotAI: Kimi K2 Thinking

Provider

Moonshot AI

4 models in catalog

Capability

JSON mode

json-mode

Catalog signal

catalog shows support

Context: 262,144

Capability probe summary

Model

moonshotai/kimi-k2-thinking

Capability

json-mode

Catalog signal

catalog shows support

Read-only check. Detection data burns after 5 minutes.

How to judge this capability

Whether MoonshotAI: Kimi K2 Thinking supports JSON mode cannot be decided from the name alone. Check model ID moonshotai/kimi-k2-thinking, provider Moonshot AI, catalog params, modality, a lightweight runtime probe, and the error body together.

Model: moonshotai/kimi-k2-thinking
Provider: Moonshot AI
Capability: JSON mode

How to prove it during checking

Read-only check. Detection data burns after 5 minutes.

Start with the catalog signal: catalog shows support, then verify with a real key in read-only mode.
The capability probe must bind model ID moonshotai/kimi-k2-thinking; do not test another model from the same provider.
JSON mode · supports frequency_penalty · supports include_reasoning · supports logit_bias · supports max_tokens

Next action

MoonshotAI: Kimi K2 Thinking JSON mode should turn into an operator action: confirm sale tags, fill the capability matrix, change route, suggest a fallback model, or mark unsupported for now.

Read-only check. Detection data burns after 5 minutes.
Catalog signal: catalog shows support
Context: 262,144

Open key checker

Read-only check. Detection data burns after 5 minutes.

Open models

MoonshotAI: Kimi K2 Thinking

Open providers

Moonshot AI