Skip to content
Model infoModelsSnapshot: 15 Oct 2026

IBM / Granite

Granite 20B Code SQL Generation

Specialized instruction-tuned adaptation of Granite 20B Code Base for generating SQL.

availableClosed weightsCoding
Official documentation (opens in a new tab)Reviewed 14 Oct 2026
7 of 17 specification fields collected
Context window8.2K8,192 tokens
Maximum outputNot collectedAwaiting source review
Input billingSee billing notesIBM watsonx.ai dedicated deployment
Output billingSee billing notesAccess and billing notes below

Specifications

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

Context / input limit
8,192 tokensSource for Context / input limit (opens in a new tab)

Combined input and output budget; not an independent maximum output allowance.

Maximum output
Not collected
Reasoning
Not collected
Tool calling
Not collected
Structured output
Not collected
Streaming
Not collected
Batch processing
Not collected
Fine-tuning
Not collected
Knowledge cutoff
Not collected
Release date
Not collected
Architecture
Decoder-only Transformer (gpt_bigcode serving architecture)Source for Architecture (opens in a new tab)
Context caching
Not collected
Published price
Dedicated model hosting; no pay-as-you-go token rate listedSource for Published price (opens in a new tab)

Hosting is billable. No numeric GPU-hour tariff was verified in this review; no per-question token price is inferred.

Access & pricing

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

IBM watsonx.ai dedicated deployment

Provider source (opens in a new tab)
API model ID
ibm/granite-20b-code-base-sql-gen-curated

Curated catalog identifier, not a drop-in shared text-generation request selector.

Access
Deploy on demand on dedicated hardware. The catalog selector provisions a deployment; inference uses that deployment’s route. Shared pay-as-you-go retirement does not imply this deployment is retired.
Regions
Product, deployment, and region restrictions apply. This record does not imply global availability or residency.
deploymentConfiguration
MAX_SEQUENCE_LENGTH=8192,DTYPE=float16Source for deploymentConfiguration (opens in a new tab)
hardware
At least one A100 or H100 GPU in the reviewed curated catalogSource for hardware (opens in a new tab)

Endpoint pricing has not been collected. Check the provider for applicable rates.

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.

IBM watsonx.ai dedicated deployment

ibm/granite-20b-code-base-sql-gen-curated
Not verified
How it works
Not collected
Minimum reusable input
Not collected
Lifetime & refresh
Not collected
How to enable it
Not collected
Usage fields
Not collected
Billing rules
Not collected
Limits & data boundaries
Not collected

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

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

IBM lists this specialized enterprise adaptation separately from public Granite 20B Code Base/Instruct. No exact matching public trained weights were found in the reviewed official repositories. Closed-weight status is inferred for this adaptation, not for its public base model.

Classification evidence (opens in a new tab)

Lifecycle dates

Dedicated catalog: available8 Aug 2024Source (opens in a new tab)

Sources reviewed