Synapse
Esc
  • Docs
    Make your first request↵
  • Streaming
  • Function calling
  • Errors and rate limits
  • Pages
    Home
  • Models
  • Pricing and cost calculator
  • Rate limits by tier
  • Enterprise, SOC 2 and data retention
  • Changelog

changelog

What shipped.

New models, API features and SDK releases. Breaking changes are announced 90 days ahead.

  1. model

    Synapse Ultra with a 1M-token context

    Our most capable model. Text, vision and code in one, priced at $4 / $16 per million tokens.

  2. model

    Synapse Code

    A 256K-context model tuned for completion and refactors. Scores 88 on our code eval.

  3. api

    Parallel tool calls

    Flash, Code, Pro and Ultra can now return several tool calls in one turn.

  4. model

    Synapse Flash 2

    Same price, 30% faster. p50 time to first token is now 120ms.

  5. platform

    ap-southeast region

    Inference in Singapore with the same endpoints and limits as us-east and eu-west.

  6. sdk

    Go SDK 1.0

    Context-aware streaming with range-over-func iterators.