# BytePlus ModelArk — https://www.byteplus.com/en/product/ModelArk # International edition of Volcano Engine (VolcEngine). # Models: 11 (6 text + 2 image + 3 video) # # ⚠️ BILLING — READ BEFORE USE # This provider points at the **standard** ModelArk inference endpoint # (`/api/v3`). Calls here are billed **per-token / per-call against your # ModelArk USD balance**. They DO NOT consume your Coding Plan # subscription quota. # # If you have a Coding Plan subscription and want calls to count against # it, use `byteplus_coding` (Anthropic protocol, friendly model # aliases) — see byteplus-coding.toml. The official BytePlus docs # explicitly warn against using `/api/v3` for Coding Plan workloads: # https://docs.byteplus.com/en/docs/ModelArk/1928261 # # Use this `byteplus` provider when you want: # - Image generation (Seedream 4.5, Seedream 5.0 Lite) # - Video generation (Seedance 1.5 Pro, Dreamina Seedance 2.0 / 2.0 Fast) # - Versioned text-model snapshots (e.g. `seed-2-0-pro-260328`) # …none of which exist on the Coding Plan endpoint. # # Pricing below is the standard real-time tier (a discounted batch tier # exists at ~half the rate; image/video billed per-call or per-token in # their own units, see comments above each model). [provider] id = "byteplus" display_name = "BytePlus ModelArk" api_key_env = "BYTEPLUS_API_KEY" base_url = "https://ark.ap-southeast.bytepluses.com/api/v3" key_required = true media_capabilities = ["image_generation", "video_generation"] # ── Text models ──────────────────────────────────────────────────────── [[models]] id = "seed-2-0-pro-260328" display_name = "Seed 2.0 Pro" tier = "frontier" context_window = 262144 max_output_tokens = 16384 input_cost_per_m = 1.00 output_cost_per_m = 6.00 supports_tools = true supports_vision = true supports_streaming = true supports_thinking = true aliases = ["seed-pro", "seed-2-pro"] [[models]] id = "seed-2-0-mini-260215" display_name = "Seed 2.0 Mini" tier = "fast" context_window = 262144 max_output_tokens = 16384 input_cost_per_m = 0.20 output_cost_per_m = 0.80 supports_tools = true supports_vision = true supports_streaming = true aliases = ["seed-mini", "seed-2-mini"] [[models]] id = "seed-2-0-code-preview-260328" display_name = "Seed 2.0 Code (preview)" tier = "smart" context_window = 262144 max_output_tokens = 16384 input_cost_per_m = 1.00 output_cost_per_m = 6.00 supports_tools = true supports_vision = false supports_streaming = true aliases = ["seed-code", "seed-2-code"] [[models]] id = "glm-4-7-251222" display_name = "GLM-4.7" tier = "smart" context_window = 204800 max_output_tokens = 16384 input_cost_per_m = 0.60 output_cost_per_m = 2.20 supports_tools = true supports_vision = false supports_streaming = true aliases = ["glm-4.7"] [[models]] id = "deepseek-v3-2-251201" display_name = "DeepSeek V3.2" tier = "smart" context_window = 131072 max_output_tokens = 8192 input_cost_per_m = 0.56 output_cost_per_m = 0.84 supports_tools = true supports_vision = false supports_streaming = true aliases = ["deepseek-v3.2"] [[models]] id = "gpt-oss-120b-250805" display_name = "GPT-OSS 120B" tier = "smart" context_window = 131072 max_output_tokens = 8192 input_cost_per_m = 0.10 output_cost_per_m = 0.50 supports_tools = true supports_vision = false supports_streaming = true aliases = ["gpt-oss-120b"] # ── Image generation models ──────────────────────────────────────────── # Billed per-piece (not per-token). Real prices in comment above each entry. # $0.0400 per image [[models]] id = "seedream-4-5-251128" display_name = "Seedream 4.5" tier = "frontier" modality = "image" input_cost_per_m = 0.0 output_cost_per_m = 0.0 supports_tools = false supports_vision = false supports_streaming = false aliases = ["seedream-4.5"] # $0.0350 per image (Dola-Seedream-5.0-lite) [[models]] id = "seedream-5-0-260128" display_name = "Seedream 5.0 Lite" tier = "frontier" modality = "image" input_cost_per_m = 0.0 output_cost_per_m = 0.0 supports_tools = false supports_vision = false supports_streaming = false aliases = ["seedream-5.0-lite", "seedream-5.0"] # ── Video generation models ──────────────────────────────────────────── # Billed per-K tokens (encoded video duration). Real per-K rates in comment. # $0.0024/K (with audio) / $0.0012/K (without audio); with-cache half the rate [[models]] id = "seedance-1-5-pro-251215" display_name = "Seedance 1.5 Pro" tier = "frontier" modality = "video" input_cost_per_m = 0.0 output_cost_per_m = 0.0 supports_tools = false supports_vision = true supports_streaming = false aliases = ["seedance-1.5-pro"] # $0.0070-$0.0077/K (input without video) / $0.0043-$0.0047/K (input with video) [[models]] id = "dreamina-seedance-2-0-260128" display_name = "Dreamina Seedance 2.0" tier = "frontier" modality = "video" input_cost_per_m = 0.0 output_cost_per_m = 0.0 supports_tools = false supports_vision = true supports_streaming = false aliases = ["dreamina-seedance-2.0"] # $0.0056/K (input without video) / $0.0033/K (input with video) [[models]] id = "dreamina-seedance-2-0-fast-260128" display_name = "Dreamina Seedance 2.0 Fast" tier = "fast" modality = "video" input_cost_per_m = 0.0 output_cost_per_m = 0.0 supports_tools = false supports_vision = true supports_streaming = false aliases = ["dreamina-seedance-2.0-fast"]