FastMetal

Models

FastMetal supports various LLM models for different use cases. View the models available on your dashboard, or query the /models endpoint.

Available Models

Models are configured by your administrator. Use the /models endpoint to list available models:

ModelDescription
mistral-voxtral-mini-3b-2507
Japan: Mistral's voxtral-mini-3b-2507
anthropic-claude-opus-4-8
Global: Claude Opus 4.8 - Most intelligent, best for agents and coding
anthropic-claude-opus-4-7
Global: Claude Opus 4.7 - Most intelligent, best for agents and coding
anthropic-claude-opus-4-6
Global: Claude Opus 4.6 - Most intelligent, best for agents and coding
anthropic-claude-sonnet-4-6
Japan: Claude Sonnet 4.6 - Best balance of speed and intelligence
anthropic-claude-haiku-4-5
Japan: Claude Haiku 4.5 - Fastest with near-frontier intelligence
minimax-m2.7
Global: Minimax's M2.7
glm-5
Global: Z.ai's GLM-5
llm-jp-3.1-8x13b-instruct4
Japan: Japanese language model optimized for instruction following
japan-gemma-4-31b
Japan: Google's Gemma 4 31B - hosted and run in Japan
japan-qwen3.6-35b
Japan: Qwen 3.6 35B-A3B - hosted and run in Japan
japan-kimi-k2.6
Japan: Moonshot AI's Kimi K2.6 - hosted and run in Japan
japan-kimi-k2.7-code
Japan: Moonshot AI's Kimi K2.7 Code - hosted and run in Japan
gpt-oss-120b
Japan: gpt-oss-120b
minimax-m3
Global: MiniMax's M3
qwen3.6-27b
Global: Qwen 3.6 27B
qwen3.8-27b
Global: Qwen 3.8 27B
kimi-k2.6
Global: Moonshot AI's Kimi K2.6
glm-4.7
Global: Z.ai's GLM 4.7
glm-4.7-flash
Global: Z.ai's GLM 4.7 Flash
glm-5.1
Global: Z.ai's GLM 5.1
mimo-v2.5
Global: Xiaomi's MiMo-V2.5
mimo-v2.5-pro
Global: Xiaomi's MiMo-V2.5-Pro
deepseek-v4-flash
Global: DeepSeek's V4 Flash
deepseek-v4-flash-0731
Global: DeepSeek's V4 Flash 0731
deepseek-v4.1-flash
Global: DeepSeek's V4.1 Flash
deepseek-v4-pro
Global: DeepSeek's V4 Pro
glm-5.2
Global: Z.ai's GLM 5.2
glm-5.3
Global: Z.ai's GLM 5.3
kimi-k3
Global: Moonshot AI's Kimi K3
qwen3.7-max
Global: Qwen 3.7 Max
qwen3.8-max
Global: Qwen 3.8 Max
grok-4.5
Global: xAI's Grok 4.5
grok-4.6
Global: xAI's Grok 4.6
gpt-5.6-sol
Global: OpenAI's GPT-5.6 Sol
gpt-5.6-terra
Global: OpenAI's GPT-5.6 Terra
gpt-5.6-luna
Global: OpenAI's GPT-5.6 Luna
gemini-3.5-flash
Global: Google's Gemini 3.5 Flash
gemini-3.7-flash
Global: Google's Gemini 3.7 Flash
inkling
Global: Thinking Machines' Inkling
anthropic-claude-opus-5
Global: Claude Opus 5 - Most intelligent, best for agents and coding
anthropic-claude-sonnet-5
Global: Claude Sonnet 5 - Best balance of speed and intelligence
anthropic-claude-fable-5
Global: Claude Fable 5 - Anthropic's flagship Mythos-class model
anthropic-claude-fable-5-1
Global: Claude Fable 5.1 - Anthropic's flagship, stronger agentic coding and long-horizon work than Fable 5
muse-glimmer-30b
Global: Meta's Muse Glimmer 30B
muse-spark-1.2
Global: Meta Muse Spark 1.2
muse-spark-1.3
Global: Meta Muse Spark 1.3
glm-5.3-flash
Global: Z.ai GLM 5.3 Flash
solar-pro4
Global: Upstage Solar Pro 4
qwen3.8-2.4t-a95b
Global: Qwen 3.8 2.4T-A95B
gpt-6-astra
Global: OpenAI's GPT-6 Astra
gpt-6-astra-pro
Global: OpenAI's GPT-6 Astra Pro
gemini-3.8-flash
Global: Google's Gemini 3.8 Flash
gemini-flash-lite-free
Global: Google's Gemini Flash Lite - free to try
mercury-2.5
Global: Inception Mercury 2.5 - diffusion LLM, the fastest reasoning model
random-free
Global: Random - free to try
nex-n2.5-mini-free
Global: Nex AGI Nex-N2.5-Mini - agentic coding model, free to try
google-nano-banana-2
Image
Global: Google Nano Banana 2 - fast, high-quality image generation
bytedance-seedream-4.5
Image
Global: Generate images
z-image-turbo
Image
Global: Generate realistic images
auto
Auto: routes each request to the model that is cheapest for its size among those supporting it (tools/vision/JSON) — billed at the resolved model's price; the listed price is what a typical small request resolves to

Model Pricing

Pricing is calculated per token. Each model has separate input and output token rates. Check the pricing page or your dashboard for current rates.

ModelInput / 1M tokensOutput / 1M tokens
mistral-voxtral-mini-3b-2507¥8.4¥8.4
anthropic-claude-opus-4-8¥893.5¥4,467.5
anthropic-claude-opus-4-7¥840¥4,200
anthropic-claude-opus-4-6¥850¥4,200
anthropic-claude-sonnet-4-6¥554.4¥2,772
anthropic-claude-haiku-4-5¥184.8¥924
minimax-m2.7¥53.61¥214.44
glm-5¥178.7¥571.84
llm-jp-3.1-8x13b-instruct4¥16¥79
japan-gemma-4-31b¥26¥101
japan-qwen3.6-35b¥32¥158
japan-kimi-k2.6¥64¥316
japan-kimi-k2.7-code¥55¥530
gpt-oss-120b¥16¥79
minimax-m3¥53.61¥214.44
qwen3.6-27b¥80.415¥482.49
qwen3.8-27b¥80.415¥571.84
kimi-k2.6¥169.765¥714.8
glm-4.7¥107.22¥393.14
glm-4.7-flash¥10.722¥71.48
glm-5.1¥250.18¥786.28
mimo-v2.5¥25.018¥50.036
mimo-v2.5-pro¥77.7345¥155.469
deepseek-v4-flash¥22.8496¥49.2146
deepseek-v4-flash-0731¥39.314¥117.942
deepseek-v4.1-flash¥50.686¥202.7441
deepseek-v4-pro¥341.317¥684.421
glm-5.2¥250.18¥786.28
glm-5.3¥250.18¥786.28
kimi-k3¥536.1¥2,680.5
qwen3.7-max¥263.5825¥790.7475
qwen3.8-max¥357.4¥1,072.2
grok-4.5¥357.4¥1,072.2
grok-4.6¥357.4¥1,072.2
gpt-5.6-sol¥893.5¥5,361
gpt-5.6-terra¥357.4¥2,144.4
gpt-5.6-luna¥35.74¥214.44
gemini-3.5-flash¥268.05¥1,608.3
gemini-3.7-flash¥134.025¥670.125
inkling¥178.7¥723.735
anthropic-claude-opus-5¥893.5¥4,467.5
anthropic-claude-sonnet-5¥357.4¥1,787
anthropic-claude-fable-5¥1,787¥8,935
anthropic-claude-fable-5-1¥1,787¥8,935
muse-glimmer-30b¥62.545¥268.05
muse-spark-1.2¥219.8968¥747.6489
muse-spark-1.3¥219.8968¥747.6489
glm-5.3-flash¥25.343¥84.4767
solar-pro4¥15.2222¥60.8888
qwen3.8-2.4t-a95b¥351.8348¥1,055.5044
gpt-6-astra¥1,717.723¥8,588.615
gpt-6-astra-pro¥1,717.723¥8,588.615
gemini-3.8-flash¥128.8292¥644.1461
gemini-flash-lite-free¥0¥0
mercury-2.5¥6.768¥25.3799
random-free¥0¥0
nex-n2.5-mini-free¥0¥0
autoresolved model's rateresolved model's rate

For full pricing details, see the pricing page.

Image Model Pricing

Image-generation models are billed per generated image rather than per token.

ModelPer image
google-nano-banana-2¥12.23
bytedance-seedream-4.5¥7.15
z-image-turbo¥1.68

Model Capabilities

Chat/Conversation — Multi-turn dialogue with context
All models support chat/conversation
Text Completion — Single-turn text generation
Text completion support