MODEL CATALOG · 2026.06

200+ Pure Models, Full Modality Coverage

All mainstream closed-source + open-source LLMs, platform directly connected to vendors or authoritative inference services. Call via OpenAI SDK, switch on demand, auto-fallback on failure.

WanFlow In-HouseAnthropicOpenAIGoogleDeepSeekMiniMaxGLMQwenSeeDanceImageVideoVoice
200+Pure Models
8Direct Vendor Connections
0MarkupSynced with Official
1SDKOpenAI Compatible
Text · Vision · Reasoning
Claude 4.8 / GPT-5 / Gemini 2.5 / DeepSeek V3.2 / Qwen 3 / GLM 4.6
Image · Video Generation
Flux 1.1 Pro · DALL-E 3 · Veo 3 · Sora · SeeDance Pro
Voice · Transcription
ElevenLabs · GPT-4o Realtime · Whisper v3 · MiniMax Audio
WanFlow In-House
Tide series trained on self-built H100 clusters, supports private deployment
IN-HOUSE · OWN STACK

WanFlow Tide
In-House Industry Models

Tide series trained on self-built H100/H200 clusters. From hardware to model, every layer is in our hands. Supports industry fine-tuning and private deployment — data never leaves your domain, performance calibrated to your business.

100%Solar Powered
RECTraceable RECs
Dedicated Deployment
H100NVIDIA Clusters
WanFlow Tide 32B
wanflow-tide-32b
SELF-HOSTEDNEW
Context
128K tok
WanFlow Tide 8B
wanflow-tide-8b
SELF-HOSTEDBASE
Context
128K tok
WanFlow Tide Code
wanflow-tide-code
SELF-HOSTEDCode
Context
128K tok
WanFlow Tide Long
wanflow-tide-long
SELF-HOSTEDLong
Context
1M tok
Claude 4 series. Opus 4.8 is the current strongest reasoning & code model; Sonnet 4.6 is the best daily driver for cost-performance.
Recommended
Claude Opus 4.8
claude-opus-4-8
NEW
Context
1M tok
Claude Sonnet 4.6
claude-sonnet-4-6
HOT
Context
200K
Claude Haiku 4.5
claude-haiku-4-5
Budget
Context
200K
Claude 3.5 Sonnet
claude-3-5-sonnet
Stable
Context
200K
GPT-5 series + o4 reasoning models. Full coverage of multimodal, vision, reasoning, and tool use.
Recommended
GPT-5
gpt-5
NEW
Context
400K
GPT-5 mini
gpt-5-mini
HOT
Context
400K
o4
o4
Reasoning
Context
200K
GPT-4o
gpt-4o
Multimodal
Context
128K
Gemini 2.5 series. 2 million context window, native multimodal, video and code understanding.
Gemini 2.5 Pro
gemini-2-5-pro
NEW
Context
2M
Gemini 2.5 Flash
gemini-2-5-flash
HOT
Context
1M
Gemini 2.5 Flash-Lite
gemini-2-5-flash-lite
Cheapest
Context
1M
Gemini 2.5 Ultra
gemini-2-5-ultra
Reasoning
Context
2M
DeepSeek-V3.2 and R1. The most cost-effective reasoning and code model combination.
Open Source
DeepSeek V3.2
deepseek-v3-2
NEW
Context
128K
DeepSeek R1
deepseek-r1
Reasoning
Context
128K
DeepSeek Coder V2
deepseek-coder-v2
Code
Context
128K
DeepSeek VL2
deepseek-vl-2
Vision
Context
32K
abab and MiniMax-M series. Long context + voice + character dialogue, fast domestic inference.
MiniMax M1
minimax-m1
NEW
Context
1M
abab 7-Preview
abab-7-preview
HOT
Context
245K
abab 6.5s
abab-6-5s
Budget
Context
245K
MiniMax Audio
minimax-audio
Voice
Context
—
GLM-4.6 series. Strong Chinese scenarios, stable tool use, domestic compliance friendly.
Open Source
GLM-4.6
glm-4-6
NEW
Context
200K
GLM-4.6-Flash
glm-4-6-flash
HOT
Context
128K
GLM-4.6V
glm-4-6v
Vision
Context
64K
GLM-4.6-Air
glm-4-6-air
Open Source
Context
128K
Qwen 3 series. Top-tier globally for Chinese scenarios, domestic compliance friendly.
Open Source
Qwen 3 Max
qwen-3-max
NEW
Context
256K
Qwen 3 235B-A22B
qwen-3-235b
MoE
Context
128K
Qwen 3 VL
qwen-3-vl
Vision
Context
128K
Qwen 3 Coder
qwen-3-coder
Code
Context
128K
Video generation model. Text-to-video / image-to-video, leading camera expression and character consistency.
Video
SeeDance 1.0 Pro
seedance-1-0-pro
NEW
Context
1080p · 10s
SeeDance 1.0 Lite
seedance-1-0-lite
HOT
Context
720p · 10s
SeeDance i2v
seedance-i2v
Image-to-Video
Context
720p · 10s
SeeDance Long
seedance-long
Long
Context
1080p · 60s

Multimodal Generation Models.

Image, video, voice — same API key, full modality coverage.

IMAGE
  • gpt-image-1 (OpenAI)
  • imagen-4 (Google)
  • minimax-image-01
  • glm-cogview-4
  • qwen-wanx-2
  • seedream-4
VIDEO
  • veo-3 (Google)
  • sora-turbo (OpenAI)
  • seedance-1-pro
  • seedance-1-lite
  • minimax-video-01
  • qwen-wanx-video
VOICE
  • gpt-4o-realtime
  • openai-tts-hd
  • whisper-large-v3
  • minimax-speech-02
  • glm-voice
  • qwen-audio

Flagship Model Comparison.

Performance data updated regularly · Last updated 2026.06

ModelVendorContextP50 LatencyCapabilities
wanflow-tide-32b
WanFlow128K0.6sIn-HousePrivate Deploy
claude-opus-4-8
Anthropic1M2.4sReasoningCode
claude-sonnet-4-6
Anthropic200K1.1sDaily Pick
gpt-5
OpenAI400K1.8sMultimodal
gemini-2-5-pro
Google2M1.5sUltra-Long Context
deepseek-v3-2
DeepSeek128K0.9sCost Effective
minimax-m1
MiniMax1M0.8sLong Context
glm-4-6
GLM200K0.7sTool Call
qwen-3-max
Qwen256K0.7sBest ChineseCompliance
seedance-1-0-pro
SeeDance10s · 1080p—Video Gen

* Latency is WANFLOW.AI US-China node P50 composite data, including TTFT. Enterprise users enjoy volume tier discounts — contact sales for details.