$5
5,000,000 credits
No bonus
Contact salesPricing
36 frontier models, one transparent rate card.
Claude Opus 4.8 · GPT-5 family · Gemini · DeepSeek V4 · Doubao · Seedream · Sora · Veo · Seedance · Nano Banana · GPT-Image · Kimi · GLM · MiniMax — all live today behind one Anthropic-compatible key. Online checkout is opening soon; activate today via sales contact or a redeem code. Failures don't bill.
DeepSeek rate card
Every DeepSeek model, one key.
V4 Pro · V4 Flash · V3.2 (OpenAI shape) · V3.2 via Anthropic Messages — all live through the same Anthropic-compatible endpoint as the rest of the gateway.
| Model | Input | Output |
|---|---|---|
| deepseek-v4-pro V4 flagship · reasoning · 64K | $0.435 ≈ ¥3.2 | $0.87 ≈ ¥6.4 |
| deepseek-v4-flash V4 fast tier · low latency | $0.14 ≈ ¥1 | $0.28 ≈ ¥2 |
| deepseek-v3-2 V3.2 chat · OpenAI shape | $0.14 ≈ ¥1 | $0.28 ≈ ¥2 |
| DeepSeek V3.2 (Anthropic Messages) V3.2 via Anthropic Messages | $0.14 ≈ ¥1 | $0.28 ≈ ¥2 |
| per 1M tokens · Illustrative CNY equivalents shown for reference; billing settles in USD. The live, nightly-refreshed rate card is at docs.bytespike.ai/pricing. | ||
How to activate
Two ways to get started today.
Online checkout is opening soon. For now, every credit top-up activates one of two ways:
Top-up credits
One credit, one US dollar.
Credits never expire. Bonus tiers kick in automatically — the dollar amount you top up determines the bonus. 1 USD = 1,000,000 credits
$50
50,000,000 credits
Bonus +1.5%
≈ $50.75 effective
Contact sales$500
500,000,000 credits
Bonus +5.0%
≈ $525 effective
Contact sales$1,250
1,250,000,000 credits
Bonus +11.2%
≈ $1,390 effective
Contact salesCredits never expire. Bonuses are applied at top-up time, not on spend. Failed requests don't bill — you can preview the cost via the x-bytespike-credits-estimated header before the body of the request consumes any credit.
DOSIA Cloud
AI agent workflows for teams.
Built on ByteSpike. SSO sign-in, admin per-user quota, one-click global / China model presets — the team layer for agent work.
Personal
Pay-as-you-go
ByteSpike credits, no minimum
Individual users running DOSIA on their own ByteSpike credits
- ByteSpike SSO sign-in
- All ByteSpike-routed models
- Personal usage view + spend cap
- DOSIA Cloud app (macOS · Windows soon)
Enterprise
Contact sales
Per-user quota · custom contract
Teams that need admin control over members, models, and spend
- Admin console for member management
- Model permission templates (global / China presets)
- Per-user usage attribution + budget caps
- Audit log + SSO/OIDC + dedicated success engineer
Learn more on /enterprise — DOSIA Cloud full capability list, use cases, architecture, security & compliance
Per-endpoint pricing
24 API surfaces, 36 models, transparent prices.
Token / per-call / per-second rates aligned to our published reference rate card. Estimated rows are flagged — the final charge always matches the estimated_credits returned at submit time, and failures don't bill.
Text — billed per million tokens
| Endpoint | Vendor | Price | Status |
|---|---|---|---|
| OpenAI Chat Completions Drop-in OpenAI Chat Completions surface. Switch base_url, keep your code; routes across GPT, Claude, Gemini, Doubao, DeepSeek, Kimi behind one key. | OpenAI compatible | $5.00 in / $30.00 out (per 1M tokens)cache read: $0.50 / 1M | live |
| OpenAI Responses OpenAI's newer Responses API surface. Stateless agentic loops, native tool use, structured outputs — all transparently supported across every model. | OpenAI compatible | $5.00 in / $30.00 out (per 1M tokens)cache read: $0.50 / 1M | live |
| Claude Messages Anthropic-shape Messages API as a first-class default. tool_use, cache_control, and thinking blocks pass through to every model that supports them; lossy-shimmed for the rest. | Anthropic compatible | $5.00 in / $25.00 out (per 1M tokens) | estimate |
| Gemini Native Google's native Gemini surface. Pass your ByteSpike key as ?key= and call generateContent / streamGenerateContent directly; routes across Gemini 3.1 Pro, Gemini 3.5 Flash, Gemini 2.5 Flash behind one key. | Google compatible | $2.00 in / $12.00 out (per 1M tokens)cache read: $0.20 / 1M | live |
Image — billed per call
| Endpoint | Vendor | Price | Status |
|---|---|---|---|
| Seedream V4 ByteDance's flagship aesthetic image generator. Strong on illustration, packaging, and editorial stills with native CJK prompt understanding. | ByteDance | $0.025 / image | estimate |
| Seedream V4.5 Refined V4 with sharper edges and tighter prompt adherence; the daily-driver upgrade. | ByteDance | $0.030 / image | estimate |
| Seedream V5lite Latency-tuned variant for product cards, thumbnails, and mass A/B image testing. | ByteDance | $0.012 / image | rolling out |
| GPT-Image-2 OpenAI's image generation endpoint. Highest prompt fidelity, strong typography handling, photorealistic outputs. | OpenAI | 1024x1024: $0.080 · 1024x1792: $0.080 · 1792x1024: $0.080 | live |
| Nano Banana Sketch-grounded image generation. Stable across iterations — pair with a quick layout pass for storyboards. | $0.018 / image | estimate | |
| Nano Banana V2 Style-leaning V2 — better at illustrative, anime, and editorial-stylized output than the base model. | $0.022 / image | estimate |
Video — billed per second
| Endpoint | Vendor | Price | Status |
|---|---|---|---|
| Sora 2 OpenAI's flagship text-to-video. Cinematic motion, strong prompt control over camera and pacing. | OpenAI | $0.100 / second | estimate |
| Sora 2 Pro Pro-tier Sora 2 with longer clips, higher temporal coherence, and finer-grained scene control. | OpenAI | 720p/1080p · per sec: $0.300 | rolling out |
| VEO 3.1 Google's flagship video generator. Notable for natural-language motion direction and integrated audio synthesis. | $0.400 / second | estimate | |
| VEO 3.1 Fast Latency-tuned VEO 3.1 for ad-hoc previews and rapid iteration on motion direction. | $0.200 / second | estimate | |
| Seedance 1.5 Pro ByteDance's mid-tier text-to-video; consistent character motion, strong on TikTok-format vertical clips. | ByteDance | $0.050 / second | estimate |
| Seedance Pro Successor to 1.5 — wider expressive range, smoother camera transitions, multi-subject scenes more reliable. | ByteDance | $0.060 / second | estimate |
| Seedance Pro Fast Pro quality, half the wall-clock. Compromises on a few prompt-control knobs, not on output resolution. | ByteDance | $0.040 / second | estimate |
| Seedance 2 4K-grade Seedance 2. Reach for it when the deliverable is a hero spot or a billboard-grade still frame. | ByteDance | $0.080 / second | estimate |
| Seedance 2 Fast Fastest Seedance 2 tier — for storyboard-grade previews where iteration speed beats fidelity. | ByteDance | $0.050 / second | estimate |
Utility — free
- BalancefreeGET /v1/balance
- Submit TaskfreePOST /v1/tasks/submit
- Query TaskfreeGET /v1/tasks/query
- Cancel TaskfreePOST /v1/tasks/cancel
Account + async-task endpoints are free to call — only the underlying generation work bills. Failures don't bill at any tier.
Estimates are conservative public-list-price approximations until that endpoint's rate goes live. The Anthropic-shape Messages endpoint covers every text model in the table — pricing is per-model, not per-protocol.
rate card refreshed · 2026-05-31
Per-model rate card
36 models, transparent prices, refreshed nightly.
We publish input / output / cache-read / cache-write per million tokens for every chat model and per-call / per-second rates for every image and video model. Sourced from the production gateway, refreshed once a day. No hidden tiers, no markup paragraphs.
Chat models
21
Image models
6
Video models
9
One key, one URL
llm.bytespike.ai/v1
All prices are quoted in U.S. dollars (USD). ByteSpike is pay-as-you-go: top up credits, spend per the rate card — no plans, no minimums, no auto-renewal. Activation today is via sales contact (sales@bytespike.ai) or redeem code at console.bytespike.ai/billing#redeem — online checkout is opening soon. Top-up credits never expire and failed requests don't bill. Stated prices exclude applicable taxes and VAT, calculated at activation based on your billing location.