Conflux Docs
Using Conflux

Models

Public model IDs users can select from Web, API, Claude CLI, and Codex CLI.

How model IDs work

Built-in provider models are exposed through the public `conflux/` namespace. The real upstream provider is intentionally hidden from end users, so users select stable IDs such as `conflux/deepseek-v4-pro` instead of provider-specific cloud names.

The exact list returned by `/v1/models` can be narrower than this documentation. Workspace policy, API key policy, user policy, disabled models, and CLI override rules can all restrict what a specific user or key is allowed to see and run.

GET https://api.yoiyoi.xyz/v1/models
Authorization: Bearer <CONFLUX_API_KEY>

System model catalog

Model IDProviderRoleContextPrice / 1M tokensCapabilitiesStatus / limitations
deepseek-v4-probd/deepseek-v4-pro-260425
bdBOTH
-safe -
In $0Out $0
TextToolsFilesConflux compat
Enabled

Custom OpenAI-compatible models stream through Conflux buffering unless a provider-specific passthrough adapter is added.

GLM 5.1bd/glm-5.1
bdBOTH
203Ksafe 192.8K
In $0Out $0
TextToolsCompletionsConflux compat
Enabled

Custom OpenAI-compatible models stream through Conflux buffering unless a provider-specific passthrough adapter is added.

DeepSeek V4 Proconflux/deepseek-v4-pro
Built-inBOTH
1.0Msafe 996.1K
In $0.435Out $0.870
TextToolsFilesWeb searchCompletions
Enabled

Runtime vision support is probed from the active Built-in route and can differ across cloud and local variants.

Gemma 4 31Bconflux/gemma4-31b
Built-inFUSE
262.1Ksafe 249.0K
In $0.130Out $0.380
TextToolsVisionFilesWeb searchCompletions
Enabled

Runtime vision support is probed from the active Built-in route and can differ across cloud and local variants.

GLM 5.1conflux/glm-5.1
Built-inBOTH
202.8Ksafe 192.6K
In $1.05Out $3.50
TextToolsFilesWeb searchCompletions
Enabled

Runtime vision support is probed from the active Built-in route and can differ across cloud and local variants.

GLM 5.2conflux/glm-5.2
Built-inBOTH
1Msafe 950K
In $0Out $0
TextToolsFilesWeb searchCompletions
Enabled

Runtime vision support is probed from the active Built-in route and can differ across cloud and local variants.

Kimi K2.5conflux/kimi-k2.5
Built-inBOTH
262.1Ksafe 249.0K
In $0.440Out $2.00
TextToolsVisionFilesWeb searchCompletions
Enabled

Runtime vision support is probed from the active Built-in route and can differ across cloud and local variants.

Kimi K2.6conflux/kimi-k2.6
Built-inBOTH
262.1Ksafe 250K
In $0.7448Out $4.655
TextToolsVisionFilesWeb searchCompletions
Enabled

Runtime vision support is probed from the active Built-in route and can differ across cloud and local variants.

Kimi K2.7 Codeconflux/kimi-k2.7-code
Built-inBOTH
262Ksafe 248.9K
In $0Out $0
TextToolsVisionFilesWeb searchCompletions
Enabled

Runtime vision support is probed from the active Built-in route and can differ across cloud and local variants.

MiniMax M2.7conflux/minimax-m2.7
Built-inSOURCE
196.6Ksafe 186.8K
In $0.300Out $1.20
TextToolsFilesWeb searchCompletions
Enabled

Runtime vision support is probed from the active Built-in route and can differ across cloud and local variants.

MiniMax M3conflux/minimax-m3
Built-inBOTH
512Ksafe 486.4K
In $0Out $0
TextToolsFilesWeb searchCompletions
Enabled

Runtime vision support is probed from the active Built-in route and can differ across cloud and local variants.

Nemotron 3 Superconflux/nemotron-3-super
Built-inBOTH
262.1Ksafe 249.0K
In $0.090Out $0.450
TextToolsFilesWeb searchCompletions
Enabled

Runtime vision support is probed from the active Built-in route and can differ across cloud and local variants.

Nemotron 3 Ultraconflux/nemotron-3-ultra
Built-inBOTH
262.1Ksafe 249.0K
In $0.600Out $3.60
TextToolsFilesWeb searchCompletions
Enabled

Runtime vision support is probed from the active Built-in route and can differ across cloud and local variants.

Qwen 3.5 397Bconflux/qwen3.5-397b
Built-inFUSE
262.1Ksafe 249.0K
In $0.390Out $2.34
TextToolsVisionFilesWeb searchCompletions
Enabled

Runtime vision support is probed from the active Built-in route and can differ across cloud and local variants.

glm-5.1glm/glm-5.1
glmBOTH
-safe -
In $0Out $0
TextToolsFilesConflux compat
Enabled

Custom OpenAI-compatible models stream through Conflux buffering unless a provider-specific passthrough adapter is added.

glm-5.2glm/glm-5.2
glmBOTH
-safe -
In $0Out $0
TextToolsFilesConflux compat
Enabled

Custom OpenAI-compatible models stream through Conflux buffering unless a provider-specific passthrough adapter is added.

GPT-5.3 Codex Highcodex/gpt-5.3-codex-high
GPTBOTH
400Ksafe 380K
In $0Out $0
TextToolsVisionFilesWeb searchConflux compat
Disabled

Disabled in the system catalog. It will not be advertised until enabled by an admin.

GPT-5.3 Codex Lowcodex/gpt-5.3-codex-low
GPTBOTH
400Ksafe 380K
In $0Out $0
TextToolsVisionFilesWeb searchConflux compat
Disabled

Disabled in the system catalog. It will not be advertised until enabled by an admin.

GPT-5.3 Codex Mediumcodex/gpt-5.3-codex-medium
GPTBOTH
400Ksafe 380K
In $0Out $0
TextToolsVisionFilesWeb searchConflux compat
Disabled

Disabled in the system catalog. It will not be advertised until enabled by an admin.

GPT-5.3 Codex Spark Highcodex/gpt-5.3-codex-spark-high
GPTBOTH
128Ksafe 121.6K
In $0Out $0
TextToolsVisionFilesWeb searchConflux compat
Disabled

Disabled in the system catalog. It will not be advertised until enabled by an admin.

GPT-5.3 Codex Spark Lowcodex/gpt-5.3-codex-spark-low
GPTBOTH
128Ksafe 121.6K
In $0Out $0
TextToolsVisionFilesWeb searchConflux compat
Disabled

Disabled in the system catalog. It will not be advertised until enabled by an admin.

GPT-5.3 Codex Spark Mediumcodex/gpt-5.3-codex-spark-medium
GPTBOTH
128Ksafe 121.6K
In $0Out $0
TextToolsVisionFilesWeb searchConflux compat
Disabled

Disabled in the system catalog. It will not be advertised until enabled by an admin.

GPT-5.3 Codex Spark XHighcodex/gpt-5.3-codex-spark-xhigh
GPTBOTH
128Ksafe 121.6K
In $0Out $0
TextToolsVisionFilesWeb searchConflux compat
Enabled

Connected-account providers require admin entitlement and active account credentials before requests can run.

GPT-5.3 Codex XHighcodex/gpt-5.3-codex-xhigh
GPTBOTH
400Ksafe 380K
In $0Out $0
TextToolsVisionFilesWeb searchConflux compat
Disabled

Disabled in the system catalog. It will not be advertised until enabled by an admin.

GPT-5.4 Highcodex/gpt-5.4-high
GPTBOTH
1.1Msafe 997.5K
In $0Out $0
TextToolsVisionFilesWeb searchConflux compat
Disabled

Disabled in the system catalog. It will not be advertised until enabled by an admin.

GPT-5.4 Lowcodex/gpt-5.4-low
GPTBOTH
1.1Msafe 997.5K
In $0Out $0
TextToolsVisionFilesWeb searchConflux compat
Disabled

Disabled in the system catalog. It will not be advertised until enabled by an admin.

GPT-5.4 Mediumcodex/gpt-5.4-medium
GPTBOTH
1.1Msafe 997.5K
In $0Out $0
TextToolsVisionFilesWeb searchConflux compat
Disabled

Disabled in the system catalog. It will not be advertised until enabled by an admin.

GPT-5.4 Mini Highcodex/gpt-5.4-mini-high
GPTBOTH
400Ksafe 380K
In $0Out $0
TextToolsVisionFilesWeb searchConflux compat
Disabled

Disabled in the system catalog. It will not be advertised until enabled by an admin.

GPT-5.4 Mini Lowcodex/gpt-5.4-mini-low
GPTBOTH
400Ksafe 380K
In $0Out $0
TextToolsVisionFilesWeb searchConflux compat
Disabled

Disabled in the system catalog. It will not be advertised until enabled by an admin.

GPT-5.4 Mini Mediumcodex/gpt-5.4-mini-medium
GPTBOTH
400Ksafe 380K
In $0Out $0
TextToolsVisionFilesWeb searchConflux compat
Disabled

Disabled in the system catalog. It will not be advertised until enabled by an admin.

GPT-5.4 Mini XHighcodex/gpt-5.4-mini-xhigh
GPTBOTH
400Ksafe 380K
In $0Out $0
TextToolsVisionFilesWeb searchConflux compat
Disabled

Disabled in the system catalog. It will not be advertised until enabled by an admin.

GPT-5.4 XHighcodex/gpt-5.4-xhigh
GPTBOTH
1.1Msafe 997.5K
In $0Out $0
TextToolsVisionFilesWeb searchConflux compat
Disabled

Disabled in the system catalog. It will not be advertised until enabled by an admin.

GPT-5.5 Highcodex/gpt-5.5-high
GPTBOTH
400Ksafe 380K
In $0Out $0
TextToolsVisionFilesWeb searchConflux compat
Enabled

Connected-account providers require admin entitlement and active account credentials before requests can run.

GPT-5.5 Lowcodex/gpt-5.5-low
GPTBOTH
400Ksafe 380K
In $0Out $0
TextToolsVisionFilesWeb searchConflux compat
Enabled

Connected-account providers require admin entitlement and active account credentials before requests can run.

GPT-5.5 Mediumcodex/gpt-5.5-medium
GPTBOTH
400Ksafe 380K
In $0Out $0
TextToolsVisionFilesWeb searchConflux compat
Enabled

Connected-account providers require admin entitlement and active account credentials before requests can run.

GPT-5.5 XHighcodex/gpt-5.5-xhigh
GPTBOTH
400Ksafe 380K
In $0Out $0
TextToolsVisionFilesWeb searchConflux compat
Enabled

Connected-account providers require admin entitlement and active account credentials before requests can run.

MiniMax-M3minimax/minimax-m3
minimaxBOTH
-safe -
In $0Out $0
TextToolsFilesConflux compat
Enabled

Custom OpenAI-compatible models stream through Conflux buffering unless a provider-specific passthrough adapter is added.

deepseek-ai/deepseek-v4-pronvidia/deepseek-v4-pro
nvidiaBOTH
-safe -
In $0Out $0
TextToolsFilesConflux compat
Enabled

Custom OpenAI-compatible models stream through Conflux buffering unless a provider-specific passthrough adapter is added.

google/gemma-4-31b-itnvidia/gemma-4-31b-it
nvidiaBOTH
-safe -
In $0Out $0
TextToolsFilesConflux compat
Enabled

Custom OpenAI-compatible models stream through Conflux buffering unless a provider-specific passthrough adapter is added.

moonshotai/kimi-k2.6nvidia/kimi-k2.6
nvidiaBOTH
-safe 250K
In $0Out $0
TextToolsFilesConflux compat
Enabled

Custom OpenAI-compatible models stream through Conflux buffering unless a provider-specific passthrough adapter is added.

nvidia/llama-3.1-nemotron-ultra-253b-v1nvidia/llama-3.1-nemotron-ultra-253b-v1
nvidiaBOTH
-safe -
In $0Out $0
TextToolsFilesConflux compat
Enabled

Custom OpenAI-compatible models stream through Conflux buffering unless a provider-specific passthrough adapter is added.

minimaxai/minimax-m2.7nvidia/minimax-m2.7
nvidiaBOTH
-safe -
In $0Out $0
TextToolsFilesConflux compat
Enabled

Custom OpenAI-compatible models stream through Conflux buffering unless a provider-specific passthrough adapter is added.

minimaxai/minimax-m3nvidia/minimax-m3
nvidiaBOTH
-safe -
In $0Out $0
TextToolsFilesConflux compat
Enabled

Custom OpenAI-compatible models stream through Conflux buffering unless a provider-specific passthrough adapter is added.

Nemotron 3 Super 120B A12Bnvidia/nemotron-3-super-120b-a12b
nvidiaBOTH
262Ksafe 248.9K
In $0Out $0
TextToolsCompletionsConflux compat
Enabled

Custom OpenAI-compatible models stream through Conflux buffering unless a provider-specific passthrough adapter is added.

Nemotron 3 Ultra 550B A55Bnvidia/nemotron-3-ultra-550b-a55b
nvidiaBOTH
262Ksafe 248.9K
In $0Out $0
TextToolsCompletionsConflux compat
Enabled

Custom OpenAI-compatible models stream through Conflux buffering unless a provider-specific passthrough adapter is added.

Qwen 3.5 397B A17Bnvidia/qwen3.5-397b-a17b
nvidiaBOTH
1Msafe 950K
In $0Out $0
TextToolsCompletionsConflux compat
Enabled

Custom OpenAI-compatible models stream through Conflux buffering unless a provider-specific passthrough adapter is added.

Google: Gemma 4 31B (free)openrouter/gemma-4-31b-it
OpenRouterBOTH
262.1Ksafe 249.0K
In $0Out $0
TextToolsVisionFiles
Enabled

OpenRouter-backed models currently stream through Conflux buffering instead of direct protocol passthrough.

NVIDIA: Nemotron 3 Super (free)openrouter/nemotron-3-super-120b-a12b
OpenRouterBOTH
1Msafe 950K
In $0Out $0
TextToolsFiles
Enabled

OpenRouter-backed models currently stream through Conflux buffering instead of direct protocol passthrough.

Selection examples

Use the same public model ID in the web selector, OpenAI-compatible requests, and Anthropic-compatible CLI requests.

{
  "model": "conflux/deepseek-v4-pro",
  "messages": [{ "role": "user", "content": "Review this API design." }]
}
claude --model conflux/minimax-m2.7 --dangerously-skip-permissions

Common limitations

Policy can hide modelsIf a model is not in `/v1/models`, it is not available to that API key or user, even if it appears in public docs.
Vision is capability-gatedImage/file support depends on the model capability matrix and provider route. Conflux may attach or normalize files, but unsupported models cannot inspect images.
Context is not unlimitedConflux advertises and enforces safe budgets below upstream context limits so responses, tools, and recovery have room to complete.
Flux costs moreFlux can run multiple source models plus a final synthesis model. Use Direct for low-latency single-model work.
Pools can hide targetsSome public model IDs may be backed by endpoint pools. Users see the stable logical model, while admin analytics show the selected provider target.