Skip to main content
One API key, one endpoint per type. Just change the model parameter.
Not sure which model to pick? Use model: "auto" — NinjaChat routes each request to the optimal model for your prompt, so you don’t have to maintain a model matrix. Smart routing guide →

Smart routing variants

These are virtual models that route to real models at runtime. Use them when you don’t want to hard-code a specific model.
IDOptimizes forBilled atGuide
autoBest model per task type (balanced)Resolved model’s rateSmart Routing →
auto-fastLowest latencyResolved model’s rateSmart Routing →
auto-cheapLowest costResolved model’s rateSmart Routing →
auto-qualityHighest qualityResolved model’s rateSmart Routing →
ensemble3-model consensus (GPT-5 + Claude Sonnet + Gemini Pro)$0.040/req flatChat →
ensemble-quality3-model consensus (GPT-5 + Claude Opus + Gemini Pro)$0.050/req flatChat →
You can also use fallback chains: "claude-opus-4.6>gpt-5>gemini-3.1-pro" — tries each model in order if the previous fails or scores below your quality threshold. See chat docs →

Pick a model

I want to…ModelCostID
Let NinjaChat decideAuto (Balanced)resolved rateauto
Build a production chatbotGPT-5$0.006gpt-5
Generate high-quality codeClaude Sonnet 4.6$0.015claude-sonnet-4.6
Do math and logical reasoningo3-mini$0.006o3-mini
Get the best possible answerClaude Opus 4.6$0.030claude-opus-4.6
Classify / extract text (budget)DeepSeek V3$0.003deepseek-v3
Fast real-time chatGemini 3 Flash$0.003gemini-3-flash
Analyze long documentsGemini 2.5 Pro$0.006gemini-2.5-pro
Highest quality (any task)Ensemble Quality$0.050ensemble-quality
Generate imagesNano Banana 2$0.05/imgnano-banana-2
Generate images (budget)FLUX.2 Klein$0.03/imgflux-2-klein
Best image qualityFLUX Kontext Max$0.10/imgflux-kontext-max
Generate videoVeo 3.1 Fast$3.00/videoveo-3.1-fast
Best video qualityVeo 3.1$5.00/videoveo-3.1
Search the webWeb Search$0.05/query

Chat models — 24 models

Full guide with examples: Chat → $0.030/req — Ultra: claude-opus-4.6 $0.015/req — Premium: claude-sonnet-4.6 $0.006/req — Standard: gpt-5 o3-mini claude-sonnet-4.5 claude-haiku-4.5 gemini-2.5-pro gemini-3-pro gemini-3.1-pro grok-4 kimi-k2 mistral-large llama-4-maverick kimi-k2.5 $0.003/req — Free tier: gpt-5-mini gemini-2.5-flash gemini-3-flash llama-4-scout deepseek-v3 qwq-32b glm-5 minimax-m2.5 ninja-1 uncensored-ai

Image models — 13 models

Full guide with examples: Images →
ModelIDCost
FLUX Kontext Maxflux-kontext-max$0.10/img
FLUX.2 Flexflux-2-flex$0.08/img
FLUX.1 Pro Ultraflux-1-pro-ultra$0.08/img
Recraft V3recraft-v3$0.08/img
Google Imagen 4google-imagen-4$0.08/img
Nano Banana Pronano-banana-pro$0.08/img
Seedreamseedream$0.08/img
FLUX.2 Proflux-2-pro$0.05/img
FLUX Kontext Proflux-kontext-pro$0.05/img
Nano Banana 2nano-banana-2$0.05/img
FLUX.1 Fillflux-1-fill$0.05/img
FLUX.2 Kleinflux-2-klein$0.03/img
Nano Banananano-banana$0.03/img

Video models — 8 models

Full guide with examples: Video →
ModelIDCost
Runway Gen-4.5runway-gen4.5$5.00/video
Veo 3.1veo-3.1$5.00/video
Runway Gen-4 Turborunway-gen4-turbo$3.00/video
Seedance 2seedance-2$3.00/video
Kling Videokling-video$3.00/video
Google Veo 2google-veo-2$3.00/video
Veo 3.1 Fastveo-3.1-fast$3.00/video
Veo 3 Fastgoogle-veo-3-fast$3.00/video

Search — $0.05/query

Full guide with examples: Search →

Live model catalog via API

Get the full model list programmatically — including capabilities, context window, speed, and recommendations:
# All models
GET https://ninjachat.ai/api/v1/models

# Filter by tier
GET https://ninjachat.ai/api/v1/models?tier=standard

# Filter by capability
GET https://ninjachat.ai/api/v1/models?capability=code

# Exclude virtual models (auto/ensemble)
GET https://ninjachat.ai/api/v1/models?virtual=false
Response includes rich metadata per model: capabilities, context_window, max_output_tokens, speed, recommended_for. Estimate costs →