Catalog

Models

Open-weight models for chat, reasoning, code, embeddings, vision and images, behind one API. Models on the same endpoint share a request shape, so switching between them is a one-string change.

Live models: 24 from 10 publishers
Regions: us-east · us-west · eu-west
Billing: Usage-based, no minimums
Dedicated GPUs: 8 models

Sort by

All 24 models

Models with price, context window, decode speed and median time to first token
Model	Task	Context	Input / 1M	Output / 1M	Speed	p50
BGE-M3BAAI · embeddings	embeddings	8K	$0.01	—	—	18 ms	→
DeepSeek R1 0528DeepSeek · reasoning	reasoning	128K	$0.70	$2.40	48 tok/s	520 ms	→
DeepSeek R1 Distill Llama 70BDeepSeek · reasoning	reasoning	128K	$0.58	$0.78	90 tok/s	230 ms	→
DeepSeek V3.1DeepSeek · chatcode	chatcode	128K	$0.55	$1.65	62 tok/s	280 ms	→
Devstral Small 2Mistral AI · code	code	128K	$0.09	$0.30	160 tok/s	120 ms	→
FLUX.1 [dev]Black Forest Labs · image	image	—	$0.025 / image		—	4.2 s	→
FLUX.1 [schnell]Black Forest Labs · image	image	—	$0.003 / image		—	900 ms	→
Gemma 3 27BGoogle · chatvision	chatvision	128K	$0.20	$0.40	140 tok/s	135 ms	→
Llama 3.1 8B InstructMeta · chat	chat	128K	$0.10	$0.14	240 tok/s	74 ms	→
Llama 3.3 70B InstructMeta · chat	chat	128K	$0.54	$0.72	96 tok/s	212 ms	→
Llama 4 MaverickMeta · chatvision	chatvision	1M	$0.27	$0.85	118 tok/s	165 ms	→
Llama 4 ScoutMeta · chatvision	chatvision	1M	$0.18	$0.59	150 tok/s	130 ms	→
Mistral Large 3Mistral AI · chatvision	chatvision	256K	$0.65	$1.60	58 tok/s	270 ms	→
Mistral Small 3.2 24BMistral AI · chatvision	chatvision	128K	$0.10	$0.25	170 tok/s	110 ms	→
Nomic Embed Text v1.5Nomic · embeddings	embeddings	8K	$0.008	—	—	14 ms	→
Phi-4 14BMicrosoft · chatreasoning	chatreasoning	16K	$0.07	$0.14	200 tok/s	85 ms	→
Qwen2.5 Coder 32B InstructAlibaba · code	code	128K	$0.28	$0.44	125 tok/s	140 ms	→
Qwen2.5 VL 72B InstructAlibaba · visionchat	visionchat	32K	$0.60	$0.80	80 tok/s	330 ms	→
Qwen3 235B A22BAlibaba · chatreasoning	chatreasoning	128K	$0.22	$0.88	85 tok/s	240 ms	→
Qwen3 32BAlibaba · chatreasoning	chatreasoning	128K	$0.30	$0.50	130 tok/s	150 ms	→
Qwen3 8BAlibaba · chat	chat	128K	$0.07	$0.12	270 tok/s	68 ms	→
Qwen3 Coder 480B A35BAlibaba · code	code	256K	$0.95	$1.90	70 tok/s	290 ms	→
Qwen3 Embedding 8BAlibaba · embeddings	embeddings	32K	$0.02	—	—	41 ms	→
Stable Diffusion 3.5 LargeStability AI · image	image	—	$0.035 / image		—	5.6 s	→

Prices in USD: token models per 1M tokens, image models per image. Price order blends input and output tokens 50/50, with image models after. p50 is the median time to first token (full response for embeddings and images) over the last 24 hours; speed is decode tokens per second for one request.