One API, all the frontier models. A single key across the Anthropic, OpenAI, and Gemini protocols, transparent USD pricing with real discounts — Claude as low as 82% off — plus cache and usage you can see.
curl https://llm.bytespike.ai/v1/chat/completions \
-H "Authorization: Bearer $BS_KEY" \
-d '{"model":"claude-sonnet-4-6","messages":[{"role":"user","content":"Hi"}]}'One gateway · 31 frontier models · 8 providers
See all →14 endpoints, one key
Image, video, text, account, async jobs — the full catalog. Each card links to the endpoint reference; bring your own integration.
1024² · text-to-image · img-to-img
OpenAI's image generation model. Highest prompt fidelity, strong typography, photorealistic output. Billed per token, discounted vs. official.
Gemini 3 Pro Image · editorial-grade
Gemini 3 Pro image generation (Nano Banana Pro). Editorial-grade output with strong prompt adherence and image-to-image editing.
Gemini 3.1 Flash Image · volume
Gemini 3.1 Flash image generation (Nano Banana 2). Fast, cheap, volume-friendly text-to-image.
Pricing is per call. Reserved balance is released on failure. estimated_credits is returned at submit time so you can show users the cost before they confirm.
Why ByteSpike
Official list price next to ours on every model — Claude as low as 82% off, the Chinese models (DeepSeek · Doubao · GLM · Kimi) at 50% off. Failed calls never bill.
31 frontier models from 8 providers — Anthropic, OpenAI, Google, DeepSeek, Doubao, GLM, Kimi, MiniMax — every model worth shipping against, behind one Anthropic-shape API.
Switch base_url, paste your key, ship. OpenAI Chat Completions and Anthropic Messages both drop in unchanged.
Lisahost LA backbone + Cloudflare edge. Median gateway overhead under 80ms. tool_use, cache_control, thinking pass through transparently.
Free balance on signup. No card. estimated_credits returned at submit time so you show users the cost before they confirm.
Self-hosted SKU, SSO/OIDC, audit log, org-level wallets, per-key quota / rate limit / IP allowlist controls. Talk to sales when you're ready.
Curated starting points
ByteSpike ships with two curated model packs so you don't pick from 30 SKUs on day one. Apply Global or China with one click; refine per team once you know which models earn their keep.
For international or English-first workloads.
Claude · GPT-5 · Gemini · DeepSeek V4
For PRC compliance or PII-sensitive workloads.
DeepSeek · Doubao · GLM · Kimi · MiniMax
Custom templates layer on top — fork a preset, swap in DeepSeek Flash, drop image gen, save as your team's stack. Per-member overrides intersect with the team allow-list server-side.
How org admins use presetsModel catalog
31 models across chat, image, and video — Anthropic, OpenAI, Google, DeepSeek, Doubao, GLM, Kimi, MiniMax, behind a single Anthropic-shape API.
Claude Fable 5
82% offAnthropic's most capable publicly released model (Mythos-class) — next-generation intelligence for long-horizon agents, native 1M context.
Claude Haiku 4.5
82% offThe fastest Claude, near-frontier intelligence — 200k context, extended thinking.
Claude Opus 4.6
82% offPrevious-gen flagship — 1M context, extended thinking. Consider migrating to Opus 4.7.
Claude Opus 4.7
82% offAnthropic's most capable general-purpose model — excels at complex reasoning and agentic coding, native 1M context.
Claude Opus 4.8
82% offAnthropic's newest flagship — further gains in complex reasoning and agentic coding, native 1M context.
Claude Sonnet 4.6
82% offThe best balance of speed and intelligence — 1M context, extended/adaptive thinking.
Claude Sonnet 5
82% offThe best balance of speed and intelligence — adaptive thinking on by default, native 1M context.
GPT-5.4
OpenAI's more economical coding and professional-work model — 1M context.
GPT-5.4 mini
OpenAI's strongest mini model — good for coding, computer use, and subagents; 400k context.
GPT-5.5
OpenAI's flagship coding and professional-work model — adjustable reasoning effort, 1M context.
GPT Image 2
20% offOpenAI's native image generation and editing (/v1/images generations + edits) — supports arbitrary resolutions.
Gemini 3.1 Flash-Lite
Low-latency, cost-efficient multimodal model — good for translation, transcription, document processing, and other high-frequency lightweight tasks.
Gemini 3.5 Flash
Google's most intelligent model — sustained frontier performance, excels at agentic and coding tasks.
Gemini Flash (latest)
Always points to the newest Gemini Flash generation; currently equivalent to gemini-3.5-flash.
Gemini Flash-Lite (latest)
Always points to the newest Gemini Flash-Lite generation; currently equivalent to gemini-3.1-flash-lite.
Gemini Pro (latest)
Google's flagship thinking model — excels at complex code, math, and STEM reasoning; 1M context.
Nano Banana Pro (Gemini 3 Pro Image)
Gemini's native image generation and editing (text-to-image / image+text edit) — supports 1K-4K resolutions and multiple aspect ratios.
Nano Banana 2 (Gemini 3.1 Flash Image)
Efficient image generation with 0.5K-4K multi-resolution output and multiple aspect ratios; integrates with search grounding.
DeepSeek V4 Flash
50% offDeepSeek's faster, more economical API — deep thinking on by default, can be disabled manually.
DeepSeek V4 Pro
50% offDeepSeek V4 Pro — significantly stronger agent capability and rich world knowledge; deep thinking on by default, can be disabled manually.
Doubao Seed 2.0 Code
50% offByteDance Doubao Seed 2.0 coding model — standout frontend generation, strong multilingual support, native multimodal vision understanding.
Doubao Seed 2.0 Lite
50% offBalances generation quality and response speed for production workloads — content creation, data analysis, and similar use cases.
Doubao Seed 2.0 Pro
50% offFlagship all-round general model, excels at complex reasoning and long-chain task execution — multimodal understanding, long context, and tool augmentation.
Doubao Seed Code
50% offDoubao coding model with precise code generation, task scheduling, and logical coordination.
GLM 5.1
50% offZhipu's next-gen flagship model — excels at code generation, long-horizon autonomous execution, and complex engineering optimization.
GLM 5.2
50% offZhipu's next-gen flagship model with native 1M-token context and adjustable thinking depth — excels at long-horizon coding and complex engineering optimization.
Kimi K2.5
50% offMoonshot Kimi K2.5 — supports text, image, and video input with thinking / non-thinking modes; strong at conversation and agent tasks.
Kimi K2.6
50% offMoonshot's strong-thinking model with multi-step tool calling and reasoning — excels at complex logical reasoning, math, and code writing.
MiniMax M2.7
Build complex agent harnesses — Agent Teams + Skills + Tools — for highly complex productivity tasks.
MiniMax M2.7
Build complex agent harnesses — Agent Teams + Skills + Tools — for highly complex productivity tasks.
MiniMax M3
50% offMiniMax's natively multimodal, 1M-context Frontier Coding model — supports text/image/video input plus thinking.
Last 24 hours
4 PRs shipped
Recent:
Feed generated:
Early customer cohort
ByteSpike just opened to public preview. The first 50 teams who integrate get a direct line to the engineering team, a retroactive adjustment on any pricing change, and their logo featured here once they're ready to be quoted publicly.
hello@bytespike.ai · response within 1 business day
FAQ
Create an account and get your API key in two minutes — or talk to sales for a private deployment.