Releasesby GlyphErrors identify their source
Before 1.2.27, the sidecar returned HTTP 502 for two failures that weren't server-side: tito_model_malformed_output (unparseable model output) and tito_upstream_error (any sampling exception, including client timeouts and dropped connections). They are replaced by: tito_sdk_parse_error (422): the output couldn't be parsed into an assistant or tool message.
Also shipped Oct 21, 2026Fireworks AI in October 2026 →
The AI.CHUNK_DOCUMENT function is a table-valued function that splits unstructured documents into...
The AI.CHUNK_DOCUMENT function is a table-valued function that splits unstructured documents into semantically coherent text chunks optimized for vector embeddings and Retrieval-Augmented Generation (RAG) workflows. This function is generally available (GA).
Automated cataloging and ingestion of Looker (Google Cloud core) metadata with Knowledge Catalog is now...
Automated cataloging and ingestion of Looker (Google Cloud core) metadata with Knowledge Catalog is now generally available (GA). This includes support for additional granular metadata (including dimension/measure definitions, default queries, and relationships) and new asset types: LookML dashboards, LookML dashboard elements, and agents.
Screen MCP tool calls and results for jailbreaks (Beta)
You can now screen MCP tool calls and their results in Unity Gateway for jailbreak and prompt injection attempts. Apply the Jailbreak guardrail on input, output, or both. See Attach a built-in policy .
[Launched] Generally Available: Azure Cosmos DB Safe Key Rotation
Safe Key Rotation for Azure Cosmos DB is now generally available. You can view timestamp metadata showing when each account key was last used, helping you identify active keys before rotating or retiring credentials.This visibility helps reduce the risk o
OpenAI models are coming to Google Cloud
OpenAI frontier models and Codex workflows are now available in Google Cloud's Model Garden via the Google Cloud Marketplace.
ChatGPT Ads expands to 26 new countries as over 100,000 advertisers have used the platform
ChatGPT Ads is now available in 26 additional countries, allowing advertisers to reach users during product discovery and research.
More from Fireworks AI platformEvery Fireworks AI platform release →
Unparseable trajectories are discarded
A parse failure makes the whole trajectory untrainable, including earlier successful turns. The cookbook RL loop drops it without retrying or stopping the run, and other rollouts continue. Parse failures don't count toward the infrastructure circuit breaker, which now classifies transient failures more accurately and reports failure counts by reason.
Failure traces are saved locally
Each failed trajectory leaves a private JSONL trace (request, prepared token IDs, raw completion, exception chain) in fireworks-tito-traces under the system temp directory, capped at 256 MiB. Traces from clean trajectories are deleted, and failures rejected by cookbook adapters use the same location and limits.
Image inputs for FiretitanSamplingClient.sample
Since 1.2.27, sample and sample_async take a keyword-only images= list of base64 data URLs for token prompts with one placeholder per image. Prompts with unprojected image chunks are rejected so sampler-owned delimiters aren't duplicated. See the cookbook's TITO failure debugging reference for every error code and how to read a trace.
Errors identify their source
Fireworks AI updated its error reporting to provide more granular source identification for failures, including new error codes for parsing and transport issues.
Unparseable trajectories are discarded
Fireworks AI now discards unparseable trajectories to prevent them from impacting training data quality.
Failure traces are saved locally
Failed model trajectories now generate private JSONL traces in the system temporary directory. These traces include request data, token IDs, and exception chains to help developers identify whether failures originated from the model, tokenizer, or renderer.
Image inputs for FiretitanSamplingClient.sample
The sample and sample_async methods now support a keyword-only images argument. This allows developers to pass base64 data URLs for token prompts using placeholders.
Action required
If you use any of the models below on serverless, migrate to a recommended replacement before November 7, 2026 . After that date, they will no longer be available via serverless endpoints.