The tables below list each model’s official direct-pay list price and integration notes. HopBase pricing is quoted one-on-one by our team based on your usage and scenario.
Claude is available in several plans (Claude Max, cloud provider direct, full-quota Max). Our team recommends the right plan and quotes pricing for your workload.
| Model ID | Official list price | Notes |
|---|---|---|
| claude-fable-5-1 Flagship | $10 / $50 | Anthropic's most capable model, cache hits $0.25/M |
| claude-fable-5 | $10 / $50 | Previous-generation Fable, cache hits $1/M |
| claude-opus-5-5 New | $4 / $20 | Next-generation Opus · 20% below Opus 5, cache hits $0.20/M |
| claude-opus-5 | $5 / $25 | Previous-generation Opus workhorse |
| claude-opus-4-8 Recommended | $5 / $25 | Top reasoning for heavy Claude Code tasks |
| claude-opus-4-7 / 4-6 | $5 / $25 | 1M-context Opus |
| claude-sonnet-5-5 New | $2 / $10 | Next-generation Sonnet · 1M context, cache hits $0.20/M |
| claude-haiku-5-5 New | $0.10 / $0.50 | Next-generation Haiku · prompts over 100K tokens billed at 5× ($0.50 / $2.50) |
| claude-sonnet-5 | $2 / $10 | Previous-generation Sonnet workhorse |
| claude-sonnet-4-6 Balanced | $3 / $15 | Balanced workhorse for long context |
| claude-sonnet-4-5-20250929 | $3 / $15 | Short alias claude-sonnet-4-5 |
| claude-sonnet-4-20250514 Deprecated / auto-routed to 4.6 | $3 / $15 | Legacy ID auto-routes to claude-sonnet-4-6 |
| claude-haiku-4-5-20251001 | $1 / $5 | Lightweight, low-cost model |
› Prices are in USD per 1 million tokens (input / output). Model availability is shown in the console.
| Model ID | Official list price | Notes |
|---|---|---|
| gpt-6-astra New | $10 / $50Cached $1 | GPT-6 flagship · phased rollout by OpenAI, at list price |
| gpt-6-sol New | $2 / $10Cached $0.2 | GPT-6 for complex coding and agent workflows · 1.05M context; $4 / $15 above 272K |
| gpt-6-luna New | $0.1 / $0.5Cached $0.01 | GPT-6 for high throughput and cost efficiency · 1.05M context; $0.2 / $0.75 above 272K |
| gpt-5.6 / -sol | $5 / $30Cached $0.5 | Next-gen flagship · 1.05M context, tiered above 272K |
| gpt-5.6-terra | $2.5 / $15Cached $0.25 | 5.6 balanced tier · 1.05M context |
| gpt-5.5 Core | $5 / $30Cached $0.5 | Complex reasoning & Codex flagship |
| gpt-5.4 | $2.5 / $15Cached $0.25 | Tiered multiplier above 272K context |
| gpt-5.4-mini | $0.75 / $4.5Cached $0.075 | Low latency, low cost |
| gpt-5.3-codex-spark | $1.75 / $14Cached $0.175 | Codex lightweight tasks |
| gpt-image-1 / 1.5 / 2 Image | $5 / $30Cached $0.5 | Image APIs billed by actual token usage |
| gpt-image-2.5-flare New Image | $5 / $30Cached $1.25 | GPT Image 2.5 everyday default · image input $8/M · five quality tiers |
| gpt-image-2.5-sunburst New Image | $5 / $30Cached $1.25 | GPT Image 2.5 higher fidelity · slower at the same settings · image input $8/M |
› Prices are in USD per 1 million tokens (input / output). Image models are billed by actual token usage.
| Model ID | Official list price | API interface | Notes |
|---|---|---|---|
| gemini-3.8-flash New | $0.75 / $3.75Cached $0.075 · $1.5 / $7.5 from 2027-01-01 | OpenAI-compatible · v1beta | Latest Flash workhorse · thinking |
| gemini-3.7-flash | $0.75 / $3.75Cached $0.075 · $1.5 / $7.5 from 2027-01-01 | v1beta only | Previous Flash workhorse |
| gemini-3.6-flash | $0.75 / $3.75Cached $0.075 · $1.5 / $7.5 from 2027-01-01 | v1beta only | Flash workhorse |
| gemini-3.5-flash | $1.5 / $9Cached $0.15 | OpenAI-compatible · v1beta | Balanced Flash tier |
| gemini-3.5-flash-lite | $0.3 / $2.5Cached $0.03 | v1beta only | Low latency, low cost |
| gemini-3.1-pro-preview / -customtools | $2 / $12Cached $0.2 | OpenAI-compatible · v1beta | Pro reasoning & long context |
| gemini-3-flash-preview | $0.5 / $3Cached $0.05 | OpenAI-compatible · v1beta | Low latency, low cost |
| gemini-3.1-flash-lite / -preview | $0.25 / $1.5Cached $0.025 | OpenAI-compatible · v1beta | Low latency, low cost |
| gemini-2.5-pro | $1.25 / $10Cached $0.125 | v1beta only | 2.5-series Pro |
| gemini-2.5-flash | $0.3 / $2.5Cached $0.03 | v1beta only | 2.5-series Flash |
| gemini-2.5-flash-lite | $0.1 / $0.4Cached $0.01 | v1beta only | Low latency, low cost |
| gemini-omni-flash-preview Video | $1.5 / $9Cached $0.15 | v1beta only | Omni multimodal · video generation |
| gemini-nano-banana-2.1 Image New | $1.5 / $7.5Cached $0.075 | OpenAI-compatible | Nano Banana 2.1 · 1K / 2K / 4K at $0.0336 / $0.0504 / $0.1134 per image |
| gemini-3-pro-image / -preview Image | $2 / $12Cached $0.2 | OpenAI-compatible · v1beta | Pro image generation |
| gemini-3.1-flash-image / -preview Image | $0.5 / $3Cached $0.05 | OpenAI-compatible · v1beta | Flash image generation |
| gemini-3.1-flash-lite-image Image | $0.25 / $1.5Cached $0.025 | OpenAI-compatible · v1beta | Flash image generation |
| gemini-2.5-flash-image Image | $0.3 / $30Cached $0.03 | OpenAI-compatible · v1beta | Nano Banana image editing |
› Prices are in USD per 1 million tokens (input / output). Cache shows the prompt-cache read rate. Available interfaces include OpenAI-compatible and native Gemini v1beta protocols; protocol type does not mean requests connect directly to Google's official API. Check the API interface column and console for per-model availability.
| Model ID | Official list price | Notes |
|---|---|---|
| glm-5.3 1M context | $1.40 / $4.40z.ai official list · Cached $0.26 | Redundant routing · thinking-length control |
| glm-5.3-flash 1M context | ¥0.80 / ¥2.80Zhipu official list · Cached ¥0.23 | 320B MoE natively multimodal · 1M context · thinking trace and tool calls |
› Prices are per 1 million tokens (input / output) in the currency shown on each row. Final charges are shown in the console. Rows priced in CNY are converted to USD and charged from your USD balance.
| Model ID | Official list price | Notes |
|---|---|---|
| kimi-k3 Flagship 1M context | $3.00 / $15.00Moonshot official list · Cached $0.30 | Latest Moonshot flagship · reasoning model · OpenAI-compatible /v1/chat/completions |
› Prices are in USD per 1 million tokens (input / output). Prompt-cache hits read at $0.30/M list. Final charges are shown in the console.
| Model ID | Official price · input / cache / output | Context | API interface |
|---|---|---|---|
| deepseek-v4.1-flash | ¥2.00 / ¥0.04 / ¥8.00 | 1M | OpenAI-compatible |
| deepseek-v4-pro-202606 | ¥8.98 / ¥0.30 / ¥26.93 | 1M | OpenAI-compatible |
| deepseek-v4-flash-202605 | ¥2.99 / ¥0.10 / ¥8.98 | 1M | OpenAI-compatible |
› Prices are in CNY per 1 million tokens and shown as input / cached input / output. DeepSeek bills peak and off-peak official rates; the table lists peak rates and off-peak is half. Charged from your USD balance after converting the official CNY list price to USD.
| Model ID | Official price · input / cache / output | Context | API interface |
|---|---|---|---|
| qwen3.8-max Flagship | ¥12 / ¥1.50 / ¥36 | 1M | OpenAI-compatible |
| qwen3.8-flash | ¥1 / ¥0.10 / ¥3 | 1M | OpenAI-compatible |
| qwen3.7-max | ¥12 / ¥2.40 / ¥36 | 1M | OpenAI-compatible |
| qwen3.7-plus | ¥2 / ¥0.40 / ¥8>256K · ¥6 / ¥1.20 / ¥24 | 1M | OpenAI-compatible |
| qwen3.7-flash | ¥0.20 / ¥0.04 / ¥0.80>32K · ¥0.60 / ¥0.12 / ¥2.40 >256K · ¥1.20 / ¥0.24 / ¥4.80 | 1M | OpenAI-compatible |
› Prices are in CNY per 1 million tokens, shown as input / cached input / output. Qwen bills in context tiers: once a request’s input crosses a threshold, the whole request is charged at that tier. Prices are the official standard-band list rates; limited-time and off-peak promotions are not reflected. Charged from your USD balance after converting the official CNY list price to USD.
| Model ID | Type and specification | Official list price |
|---|---|---|
| grok-4.6 | Chat · 500K context · per 1M tokens | $2 / $6Cache read $0.50 |
| grok-4.5 | Chat · 500K context · per 1M tokens | $2 / $6Cache read $0.30 |
| grok-4.3 | Chat · 1M context · per 1M tokens | $1.25 / $2.50Cache read $0.20 |
| grok-4.20-0309-reasoning | Reasoning · 2M context · per 1M tokens | $1.25 / $2.50Cache read $0.20 |
| grok-4.20-multi-agent-0309 | Multi-agent · 2M context · per 1M tokens | $1.25 / $2.50Cache read $0.20 |
| grok-imagine-image Image | Image · 1k · per image | $0.02Input image $0.002 each |
| grok-imagine-image-2.0 Image | Image · 1k / 2k · per image | $0.04 / $0.06Input image $0.03 each |
| grok-imagine-image-quality Image | Image · 1k / 2k · per image | $0.05 / $0.07Input image $0.01 each |
| grok-imagine-video-1.5 Video | Video · 480p / 720p / 1080p · per output second | $0.08 / $0.14 / $0.25Reference image $0.01 each |
› Chat prices are in USD per 1 million tokens, shown as input / output. A request whose input reaches 200K tokens is charged at double rates for the whole request. Image and video models are billed per image and per output second, not by token usage. Video sits in its own plan group and uses the asynchronous task API.
Seedance · Video
| Tier | Model ID | Resolution | No video input | With video input |
|---|---|---|---|---|
| SD2.5 | dreamina-seedance-2-5-260628 | 480p / 720p integer 4-30s / auto | $10.70 | $6.40 |
| 1080p | $11.70 | $7.00 | ||
| Standard | dreamina-seedance-2-0-260128 | 480p / 720p | $7.00 | $4.30 |
| 1080p | $7.70 | $4.70 | ||
| 4k | $4.00 | $2.40 | ||
| Fast | dreamina-seedance-2-0-fast-260128 | 480p / 720p | $5.60 | $3.30 |
| Mini | dreamina-seedance-2-0-mini-260615 | 480p / 720p | $3.50 | $2.10 |
Seedream · Image
| Model ID | Resolution | Official list price | Notes |
|---|---|---|---|
| seedream-5-0-pro Recommended | 1K / 2K · PNG / JPEG · standard / fast | $0.045 / $0.09≤2.36MP / >2.36MP | Text-to-image / image-to-image / partial edits |
| seedream-5-0-lite | 2K / 3K / 4K · PNG / JPEG · standard | $0.035 | Seedream 5.0 Lite |
| seedream-4-5 | 2K / 4K · JPEG · standard | $0.04 | Previous-gen image workhorse |
› Video prices are in USD per 1 million video tokens at the official list. Usage increases with resolution, frame rate, and duration. HopBase-hosted result URLs (outputs and source_outputs) are valid for 24 hours from task completion. Image prices are per generated image.
| Model ID | Representative specification | Official list price |
|---|---|---|
| kling-v3-omni | 720p, silent, no reference | ¥0.60 / s |
| kling-v3-omni | 1080p with audio | ¥1.00 / s |
| kling-v3-omni | 720p feature reference | ¥0.90 / s |
| kling-v3 | 720p, silent | ¥0.60 / s |
| kling-v3 | 1080p with audio | ¥1.20 / s |
| kling-v3-motion-control | 1080p motion control | ¥1.20 / s |
| kling-image-v3-omni | 1K / 2K output | ¥0.20 / image |
| kling-image-v3-omni | 4K output | ¥0.40 / image |
› Official list prices. Exact available buckets and account settlement follow the signed-in console and actual task charge. Charged from your USD balance after converting the official CNY list price to USD.
| Model ID | Representative specification | Official list price |
|---|---|---|
| midjourney-v8-2 | Standard render · 4 images per task | $0.0067 / image |
| midjourney-v8-2 | --hd 2K render · 4 images per task | $0.0108 / image |
› Official list prices, billed per returned image. Every task returns four images.
| Model ID | Representative specification | Official list price |
|---|---|---|
| MiniMax-H3 | 768P output · 4–15s per clip | $0.08 / s |
| MiniMax-H3 | 2K output · 4–15s per clip | $0.13 / s |
| MiniMax-H3 | Reference video input, per input second (matches output tier) | $0.08–0.13 / s |
| MiniMax-H3 | Input images — first 5 free, then per image | $0.04 / image |
| MiniMax-H3-Max New | 480P output · 5–15s per clip · fast generation | $0.05 / s |
| MiniMax-H3-Max | 768P output · 5–15s per clip · fast generation | $0.08 / s |
| speech-2.8-hd New | Text-to-speech · HD voice quality · OpenAI /v1/audio/speech & native t2a_v2 · billed per character, Chinese characters count as 2 | $100 / 1M chars |
| speech-2.8-turbo New | Text-to-speech · Turbo, lower latency · same two endpoints · billed per character, Chinese characters count as 2 | $60 / 1M chars |
| Voice cloning New | Rapid voice cloning · charged once on the voice's first synthesis; cloning and upload are free · preview billed as TTS characters | $1.5 / voice |
› Video: failed or safety-filtered generations are never billed; audio reference input is free; billing follows actual usage on completed clips only. Speech: billed on the input text — every character counts 1 and each Chinese character counts 2 (punctuation, spaces and emoji count 1); /v1/audio/speech accepts up to 4,096 characters, native t2a_v2 up to 10,000 with streaming.
| Model ID | Representative specification | Official list price |
|---|---|---|
| wan3.0-video | 480P output · 2–30s per clip | ¥0.30 / s |
| wan3.0-video | 720P output · 2–30s per clip | ¥0.60 / s |
| wan3.0-video | 1080P output · 2–30s per clip | ¥1.20 / s |
| wan3.0-video | Reference video input, per input second (matches output tier) | ¥0.30–1.20 / s |
| happyhorse-1.1-t2v / -i2v / -r2v | 480P output · 3–15s per clip | ¥0.45 / s |
| happyhorse-1.1-t2v / -i2v / -r2v | 720P output · 3–15s per clip | ¥0.90 / s |
| happyhorse-1.1-t2v / -i2v / -r2v | 1080P output · 3–15s per clip | ¥1.20 / s |
› Official list prices. Failed generations are never billed — settlement follows actual usage on completed clips only. Wan 3.0 also bills reference-video input seconds at the output tier; HappyHorse bills output seconds only. Charged from your USD balance after converting the official CNY list price to USD.
