Releasesby IbisErrors identify their source
Update2 days agoUnverifiedAdded Oct 22, 2026
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.
Topics: Streaming, Developer tools
More Fireworks AI platform releases
Every Fireworks AI platform releaseUnparseable 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.
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.
Recommended migrations
- GPT OSS 120B , move to a dedicated deployment of GPT OSS 120B Inkling , migrate to Muse Spark Open (US endpoint, 50% US regional premium waived) NVIDIA Nemotron 3 Ultra , move to a dedicated deployment of Nemotron Super VL NVIDIA Nemotron 3.5 Lightning 30B A3B , move to a dedicated deployment of Nemotron Super VL MiniMax M3 (US endpoint only) , migrate to...
New models on Serverless
- DeepSeek V4.1 Flash 1017: accounts/fireworks/models/deepseek-v4p1-flash-1017 DeepSeek V4.1 Flash 1017 (US endpoint): accounts/fireworks/models/deepseek-v4p1-flash-1017-us Muse Spark Open (US endpoint, 50% US regional premium waived): accounts/fireworks/models/muse-spark-open MiniMax M3.1: coming to serverless after its open-source release, expected...
Training Upcoming Serverless Training deprecation: DeepSeek V4 Flash (0731) On November 2, 2026 , DeepSeek V4...
Training Upcoming Serverless Training deprecation: DeepSeek V4 Flash (0731) On November 2, 2026 , DeepSeek V4 Flash (0731) will be removed from the shared Serverless Training pool. This change applies only to Serverless Training. If you train DeepSeek V4 Flash (0731) on Serverless Training, migrate to DeepSeek V4.1 Flash 1017 before November 2, 2026.
Training New Serverless Training model: Muse Spark Open Muse Spark Open is now available for LoRA workloads...
Training New Serverless Training model: Muse Spark Open Muse Spark Open is now available for LoRA workloads on the shared Serverless Training pool, with up to 262K context. Use accounts/fireworks/models/muse-spark-open-nvfp4 as the base model. See the Serverless Training guide for setup and the training model catalog for current availability.
Also shipped on Oct 22, 2026
Fireworks AI in October 2026Trace and evaluate Cursor agents with Langfuse, powered by ClickHouse
Users can now connect Cursor to Langfuse to trace agent turns and analyze model and tool calls, with ClickHouse providing the underlying analytics infrastructure.
The AI.TRANSLATE function lets you translate text strings between languages
The AI.TRANSLATE function lets you translate text strings between languages. AI.TRANSLATE supports standard Neural Machine Translation (NMT) workloads and Translation LLM (TLLM) capabilities. This function is generally available (GA).
Replit Monitoring Agent: Your very own software engineering team
Replit Monitoring Agent is an autonomous tool that investigates production alerts, identifies causes, applies fixes, and deploys them to production for internal apps.
Microsoft Entra ID connector is available in Beta
The managed Microsoft Entra ID connector is now available in Beta in Databricks Lakeflow Connect. Use the connector to ingest Microsoft Entra ID users, groups, applications, service principals, directory roles, identity configuration, and audit and sign-in logs into Databricks.
You can configure Direct VPC ingress for Cloud Run instances ( Preview )
You can configure Direct VPC ingress for Cloud Run instances ( Preview ). Direct VPC ingress lets Cloud Run instances on a VPC network use their own ephemeral private IP address to communicate with each other and with other resources on the VPC network.
Gemini Enterprise: Single-user quota limits
Gemini Enterprise administrators can configure single-user quota limits on the Usage & Spending > Usage page in the Google Cloud console to prevent an individual user from using all of an edition's shared subscription quota pool.