Skip to content

We raised $7.5m to build the intelligence layer for AI

Sign in Get an API key

Features

Spend controlsBudgets and limits for every key and teammate
GuardrailsBlock risky requests before a model sees them.
Usage analyticsSpend, speed and failures across your requests
Request logsEvery request with its cost and outcome
FallbacksBackup models, tried in the order you set
Bring your own keyYour provider keys, with limits and tracking
PlaygroundTest prompts on up to three models
IntegrationsCoding agents, SDKs and frameworks

Models

CatalogEvery model you can call, with its price
BenchmarksIntelligence, cost and speed, compared by task
RankingsUsage share and the fastest, most reliable routes

Use cases

Agent buildersComing soon
AI startupsComing soon
AgenciesComing soon
Enterprise teamsComing soon

Resources

BlogComing soon
Case studiesComing soon
DiscordAsk the team and other developers

Qwen3.8 Max

fireworks-ai/qwen3p8-max

Tools
JSON mode
Vision
Streaming

Released Aug 5, 2026

Context: 262.1K
Max out: 131K
From / 1M: $2.00
Intelligence: 40.2

Providers

Input / 1MOutput / 1MCache read / 1MSpeedFirst token

Fireworks AI
fireworks-ai/qwen3p8-max
Input / 1M$2.00
Output / 1M$6.00
Cache read / 1M$0.25
Speed106.3 t/s
First token485 ms

Hopscotch is an LLM gateway: one API for top models, with spend limits and failover built in.

© 2026 Hopscotch

Product

Spend controls
Guardrails
Usage analytics
Request logs
Fallbacks
Bring your own key
Playground
Integrations
Pricing

Models

All models
Benchmarks
Rankings
OpenAI
Google
Alibaba
Zhipu AI
Anthropic
DeepSeek
Meta
MiniMax

Developers

Docs
Quickstart
API reference
llms.txt

Company

Team
Contact
Discord

Legal

Privacy Policy
Terms of Service
Service Level Agreement

AnthropicCohereDeepInfraDeepSeekFireworks AIGoogleMetaMiniMaxMoonshot AIOpenAIZ.ai

hopscotch