Skip to content
Model infoModelsSnapshot: 15 Oct 2026

WRITER / Palmyra

Palmyra X4

Language tasks and workflow automation.

availableClosed weightsAgentsDocument analysis
Official documentation (opens in a new tab)Reviewed 14 Oct 2026
10 of 18 specification fields collected
Context window128K128,000 tokens
Maximum output4.1K4,096 tokens
Input / 1M tokens$2.50WRITER API
Output / 1M tokens$10.00USD · observed 14 Oct 2026

Specifications

Provider-documented facts. “Not collected” means we have not verified that field yet, not that the feature is unsupported.

Reasoning
Not collected
Tool calling
SupportedSource for Tool calling (opens in a new tab)

The provider explicitly supports tool calling for Palmyra X4 and later models. Built-in product actions are separate from native model output.

Structured output
SupportedSource for Structured output (opens in a new tab)

The guide explicitly names X4 and X5 for JSON Schema response_format. No unsupported flag is inferred for unnamed models.

Streaming
SupportedSource for Streaming (opens in a new tab)

Chat stream=true returns server-sent events.

Batch processing
Not collected
Fine-tuning
Not collected
Knowledge cutoff
Not collected
Release date
2024-10-09Source for Release date (opens in a new tab)

Original Palmyra X 004 release named in the dated product update; current API alias is X4.

Parameters
Not collected
Architecture
Not collected
Languages
Not collected
Context caching
Not collected

The chat usage schema includes cached_tokens, but support, rates, TTL, thresholds and configuration for this exact model are unverified.

Generation controls
Chat accepts max_tokens, temperature (default 1), top_p, n, stop, logprobs, stream, tools and tool_choice.Source for Generation controls (opens in a new tab)

The max_tokens ceiling is model-specific. No undocumented ranges, top_p default or reasoning-effort control are inferred.

Release identity
Commercial API selector palmyra-x4Source for Release identity (opens in a new tab)

Model names and hosted release aliases do not identify immutable weights. Product orchestration can invoke other models and tools.

Access & pricing

Prices belong to an endpoint and billing tier. A provider listing does not guarantee access for every account or region.

API model ID
palmyra-x4
Access
WRITER API account/key required. API, Agent Builder and no-code share the published model token rates. WRITER Agent has a distinct X5 rate and separately billed tools.
Regions
Not collected
Tool calling
SupportedSource for Tool calling (opens in a new tab)

The provider explicitly supports tool calling for Palmyra X4 and later models. Built-in product actions are separate from native model output.

Structured output
SupportedSource for Structured output (opens in a new tab)

The guide explicitly names X4 and X5 for JSON Schema response_format. No unsupported flag is inferred for unnamed models.

Streaming
SupportedSource for Streaming (opens in a new tab)

Chat stream=true returns server-sent events.

INPUT / 1M TOKENS$2.50
OUTPUT / 1M TOKENS$10.00

Direct WRITER API / Agent Builder / no-code token rates. Tools, Knowledge Graph and other product usage are billed separately; cached-input pricing is unverified. Cached input price: not collected.

USD · observed 14 Oct 2026 · Pricing evidence (opens in a new tab)

Data & enterprise terms

Retention, training use, regional processing and enterprise terms have not been summarized for this endpoint.

Caching

Caching can reuse earlier input processing. Availability, lifetime and billing depend on the endpoint. A cache hit still generates a new response.

01 / INITIAL INPUTProcess & retain

The system processes eligible input and retains reusable state.

02 / LATER REQUESTCheck available state

Reuse depends on matching input, session state and the cache policy.

03 / CACHE HITReuse & generate

Reusing state can reduce repeated work. Any charges follow the billing rules below.

WRITER API

palmyra-x4
Not verified

No model-specific provider cache contract or price verified. A cached_tokens usage field alone does not establish cache eligibility, lifetime or a discounted rate.

How it works
Not collected

No model-specific provider cache contract or price verified. A cached_tokens usage field alone does not establish cache eligibility, lifetime or a discounted rate.

Minimum reusable input
Not collected

No model-specific provider cache contract or price verified. A cached_tokens usage field alone does not establish cache eligibility, lifetime or a discounted rate.

Lifetime & refresh
Not collected

No model-specific provider cache contract or price verified. A cached_tokens usage field alone does not establish cache eligibility, lifetime or a discounted rate.

How to enable it
Not collected

No model-specific provider cache contract or price verified. A cached_tokens usage field alone does not establish cache eligibility, lifetime or a discounted rate.

Usage fields
Not collected

No model-specific provider cache contract or price verified. A cached_tokens usage field alone does not establish cache eligibility, lifetime or a discounted rate.

Billing rules
Not collected

No model-specific provider cache contract or price verified. A cached_tokens usage field alone does not establish cache eligibility, lifetime or a discounted rate.

Limits & data boundaries
Not collected

No model-specific provider cache contract or price verified. A cached_tokens usage field alone does not establish cache eligibility, lifetime or a discounted rate.

No verified cache rate for this endpoint. An empty rate is not a zero charge.

Caching reviewed 14 Oct 2026

What the measurements say

No benchmark observations have been added for this model. We do not turn a vendor description into a performance score.

Latency and throughput depend on the host, request size, reasoning settings and load. There is no measured speed or global usage figure in this record.

The record behind the model

Why it is included

Hosted-release distribution inference: the reviewed exact API model is commercially served by WRITER and no public trained weights for this exact release were identified. Related X-4.3 and explicitly dummy-weight repositories do not establish that the current X4 selector serves those same downloadable weights. Exact checkpoint mapping remains under review.

Classification evidence (opens in a new tab)

Lifecycle dates

Planned deprecation14 Dec 2026Source (opens in a new tab)

Sources reviewed