gpt-oss local model profile

Evidence snapshot: 2026-09-06

gpt-oss is a local model family available through Ollama. The published local profile lists 20B / 120B variants, 128K context, and a typical 14-65GB download range; effective speed and capacity depend on hardware and quantization.

gpt-oss is a local model family available through Ollama. It is intended for agentic tasks, structured output, code, and configurable reasoning, with 20B / 120B variants and a typical 14-65GB download range. Effective context, speed, and quality depend on the selected quantization and hardware.

gpt-oss runs on your own hardware through Ollama in Kendr Desktop, so the trade it asks you to make is machine capacity against privacy and per-token cost, not price per million.

gpt-oss is documented as a local Ollama family. This page does not claim a Kendr-hosted API route, hosted availability, a Kendr markup, or a routing receipt.

gpt-oss accepts text and returns text. The snapshot records text, tools, and reasoning as capabilities or interfaces.

$0 token fee. Hardware, storage, memory, electricity, and operational costs remain user-provided.

No comparable third-party benchmark value is present in the 2026-09-06 snapshot. Missing values are not estimated.

No comparable popularity rank or traffic share is present in the 2026-09-06 snapshot. No comparable routing rank or traffic-share observation is attached to this dated profile; missing adoption evidence is not estimated.

Profile data was reviewed for the 2026-09-06 snapshot. Claims retain their source dates, and unavailable fields are shown as unavailable instead of estimated.

Profile facts

Profile type: Local model family
Provider or publisher: Open-weight / Ollama
Context window: 128K
Snapshot date: 2026-09-06
Knowledge cutoff: Not disclosed

When to pick gpt-oss

gpt-oss runs on your own hardware through Ollama in Kendr Desktop, so the trade it asks you to make is machine capacity against privacy and per-token cost, not price per million.

Each condition below is derived from this model's own values in the 2026-09-06 snapshot, compared only against models publishing the same field.

Reach for it when the data cannot leave your machine. This family runs locally through Ollama in Kendr Desktop, at a typical download of 14-65GB. There is no token fee and no cloud round trip, and the cost is the hardware you run it on.
Reach for it when the model has to call tools. Tool calling is recorded in the snapshot, so this route can drive an agent loop rather than only answer in prose.

Context, modalities, and identity

gpt-oss accepts text and returns text. The snapshot records text, tools, and reasoning as capabilities or interfaces.

Provider or publisher: Open-weight / Ollama
Context window: 128K
Maximum output: No verified numeric limit
Input modalities: text
Output modalities: text
Knowledge cutoff: Not disclosed

Ollama variants and hardware boundary

The documented Ollama family uses gpt-oss. It is suited to Agentic tasks, structured output, code, and configurable reasoning. Actual throughput and maximum usable context vary with the selected variant, quantization, runtime, RAM, and VRAM.

Ollama identifier: gpt-oss
Variants: 20B / 120B
Typical download: 14-65GB

Model overview

Reference model ID: Not separately documented
Canonical version: Not separately documented
Hugging Face ID: Not separately documented

Capabilities and supported controls

Capabilities and parameters are reported from the dated catalog and source records; their presence does not guarantee identical behavior across every provider route.

Supported parameters: No verified parameter list

text
tools
reasoning

Local operating cost

$0 token fee. Hardware, storage, memory, electricity, and operational costs remain user-provided.

Local inference is not a zero-cost operation: the user supplies compute, memory, storage, electricity, and maintenance.

Price date: 2026-09-06
Input: $0 per 1M tokens
Cached input: No verified rate
Output: $0 per 1M tokens

Benchmark evidence and limitations

No comparable third-party benchmark value is present in the 2026-09-06 snapshot. Missing values are not estimated.

Local performance varies by exact variant, quantization, hardware, runtime, and settings.

Popularity and market context

Third-party catalog rank: No verified rank
Global market share: Not inferred
Observation date: 2026-09-06

Frequently asked questions

When should I use gpt-oss?

gpt-oss runs on your own hardware through Ollama in Kendr Desktop, so the trade it asks you to make is machine capacity against privacy and per-token cost, not price per million. Reach for it when the data cannot leave your machine: This family runs locally through Ollama in Kendr Desktop, at a typical download of 14-65GB. There is no token fee and no cloud round trip, and the cost is the hardware you run it on. Reach for it when the model has to call tools: Tool calling is recorded in the snapshot, so this route can drive an agent loop rather than only answer in prose.

What is gpt-oss?

What context window does gpt-oss have?

The dated profile lists 128K of context. Provider routes, variants, and runtime configuration can impose lower effective limits.

What benchmark evidence is available for gpt-oss?

No comparable third-party benchmark value is present in the 2026-09-06 snapshot. Missing values are not estimated. Local performance varies by exact variant, quantization, hardware, runtime, and settings.

Can gpt-oss run locally?

gpt-oss is documented here as an Ollama family with the identifier gpt-oss. Hardware, quantization, context settings, and the selected variant determine practical speed and memory use.

Sources and evidence dates

gpt-oss on Ollama (primary, checked 2026-09-06)