Back to the service list
Model provider / Model API

Logfare

An OpenAI-compatible API offering several models free of charge. Prompts and responses are recorded, and consenting to model training or data sale changes which models and limits you can use.

Platforms: API

Plans and what they cover

Not disclosed means the official page states no figure; not researched means we have not checked the item yet. Neither is the same as "no limit".

Standard

Free

Free

Free

An account and membership of the Discord server are required. Prompts, responses and metadata are recorded. You can object to use for internal evaluation data. Model training and data sale are separate opt-ins. A temporary limit increase was announced until 2026-11-14 23:27 UTC. The figures above are the normal ones.

Models and their allowances

Listed models are those we could confirm. When a model has no individual figure, check the shared plan allowance or separate contract terms as well.

  • Qwen 3.8 27BAvailable

    Included in Standard. Uses the plan’s shared allowance. The model list changes over time.

    This model’s own allowance: not researched, or shared across the plan. See the usage limits below.

  • Gemma 4 26BAvailable

    In the public model API this model is on Standard and is exempt from the account and network request caps. Concurrency and provider-side limits still apply.

    This model’s own allowance: not researched, or shared across the plan. See the usage limits below.

  • DeepSeek V4.1 Flash (Economy)Available

    In the public model API this model is on Standard and is exempt from the account and network request caps. Concurrency and provider-side limits still apply. It has a different ID from the standard version.

    This model’s own allowance: not researched, or shared across the plan. See the usage limits below.

  • GLM-5.3-FlashNot available

    On Premium in the model API at the time of checking. Requires consenting to model training.

    This model’s own allowance: not researched, or shared across the plan. See the usage limits below.

Usage limit details

Requests, standard account40 requests / minute
Scope
Whole plan
Window
minute
Resets
Fixed UTC windows
On overage
On 429/503, wait at least the Retry-After time
Rollover
Not researched

Normally 40 requests per minute, shared across all covered models, APIs and apps. For the first 24 hours it is 10 per minute. Separate limits apply per IP, per provider and per model. Temporary limit increases are not included.

Official source for this limit
Requests, standard account1,000 requests / hour
Scope
Whole plan
Window
hour
Resets
Fixed UTC windows
On overage
On 429/503, wait at least the Retry-After time
Rollover
Not researched

The hourly cap for the same standard account. For the first 24 hours it is 200. It is assessed separately from the per-minute, per-day and concurrency caps.

Official source for this limit
Requests, standard account5,000 requests / day
Scope
Whole plan
Window
day
Resets
Fixed UTC windows
On overage
On 429/503, wait at least the Retry-After time
Rollover
Not researched

The daily cap for the same standard account. For the first 24 hours it is 500. It is assessed separately from the per-minute, hourly and concurrency caps.

Official source for this limit
Concurrent requests, standard account6 concurrent / concurrent
Scope
Whole plan
Window
concurrent
Resets
The slot frees up when the request finishes
On overage
429 after waiting for a free slot and timing out
Rollover
Not applicable

For the first 24 hours it is 2 concurrent. Per-account settings and an overall cap also apply.

Official source for this limit

Officially confirmed · Checked

Official plan page for the latest terms
Official sources read (3)

Premium (consent to training use)

Free

Free

Free

Free of charge, but Premium models require consenting to use your content for model training. Prompts and responses are recorded, and content already trained on may not be removable from the weights after you withdraw consent. The first 24 hours use a new-account allowance. Temporary limit increases are not included in the normal figures.

Models and their allowances

Listed models are those we could confirm. When a model has no individual figure, check the shared plan allowance or separate contract terms as well.

  • Qwen 3.8 27BAvailable

    Standard models are also available.

    This model’s own allowance: not researched, or shared across the plan. See the usage limits below.

  • Gemma 4 26BAvailable

    On Standard. Exempt from the request cap, but concurrency and provider-side limits still apply.

    This model’s own allowance: not researched, or shared across the plan. See the usage limits below.

  • DeepSeek V4.1 Flash (Economy)Available

    On Standard. Exempt from the request cap, but concurrency and provider-side limits still apply.

    This model’s own allowance: not researched, or shared across the plan. See the usage limits below.

  • Kimi K2.7 CodeConditional

    On Premium in the public model API. Requires consenting to use for model training.

    This model’s own allowance: not researched, or shared across the plan. See the usage limits below.

  • GLM 5.2Conditional

    On Premium in the public model API. Requires consenting to use for model training.

    This model’s own allowance: not researched, or shared across the plan. See the usage limits below.

  • GLM-5.3-FlashConditional

    Requires consenting to model training. The API model list changes without notice.

    This model’s own allowance: not researched, or shared across the plan. See the usage limits below.

Usage limit details

Requests, standard account80 requests / minute
Scope
Whole plan
Window
minute
Resets
Fixed UTC windows
On overage
On 429/503, wait at least the Retry-After time
Rollover
Not researched

Normally 80 requests per minute, shared across all covered models. For the first 24 hours it is 10 per minute. IP and model-side limits also apply. Temporary limit increases are not included.

Official source for this limit
Requests, standard account2,000 requests / hour
Scope
Whole plan
Window
hour
Resets
Fixed UTC windows
On overage
On 429/503, wait at least the Retry-After time
Rollover
Not researched

The hourly cap for Premium. For the first 24 hours it is 200.

Official source for this limit
Requests, standard account10,000 requests / day
Scope
Whole plan
Window
day
Resets
Fixed UTC windows
On overage
On 429/503, wait at least the Retry-After time
Rollover
Not researched

The daily cap for Premium. For the first 24 hours it is 500.

Official source for this limit
Concurrent requests, standard account12 concurrent / concurrent
Scope
Whole plan
Window
concurrent
Resets
The slot frees up when the request finishes
On overage
429 after waiting for a free slot and timing out
Rollover
Not applicable

Concurrency for Premium. For the first 24 hours it is 2.

Official source for this limit

Officially confirmed · Checked

Official plan page for the latest terms
Official sources read (3)

Pro (consent to training and data sale)

Free

Free

Free

Free of charge, but Pro requires consenting to model training plus a separate opt-in to sell or license your content to third parties. Content already sold cannot be recovered after you withdraw. Prompts and responses are recorded. The first 24 hours use a new-account allowance. Temporary limit increases are not included in the normal figures.

Models and their allowances

Listed models are those we could confirm. When a model has no individual figure, check the shared plan allowance or separate contract terms as well.

  • Qwen 3.8 27BAvailable

    Standard models are also available.

    This model’s own allowance: not researched, or shared across the plan. See the usage limits below.

  • Gemma 4 26BAvailable

    On Standard. Exempt from the request cap, but concurrency and provider-side limits still apply.

    This model’s own allowance: not researched, or shared across the plan. See the usage limits below.

  • DeepSeek V4.1 Flash (Economy)Available

    On Standard. Exempt from the request cap, but concurrency and provider-side limits still apply.

    This model’s own allowance: not researched, or shared across the plan. See the usage limits below.

  • Kimi K2.7 CodeConditional

    On Premium in the public model API. Requires consenting to use for model training.

    This model’s own allowance: not researched, or shared across the plan. See the usage limits below.

  • GLM 5.2Conditional

    On Premium in the public model API. Requires consenting to use for model training.

    This model’s own allowance: not researched, or shared across the plan. See the usage limits below.

  • GLM-5.3-FlashConditional

    A Premium model. Requires consenting to training use.

    This model’s own allowance: not researched, or shared across the plan. See the usage limits below.

Usage limit details

Requests, standard account160 requests / minute
Scope
Whole plan
Window
minute
Resets
Fixed UTC windows
On overage
On 429/503, wait at least the Retry-After time
Rollover
Not researched

Normally 160 requests per minute. For the first 24 hours it is 10 per minute. IP and model-side limits also apply. Temporary limit increases are not included.

Official source for this limit
Requests, standard account4,000 requests / hour
Scope
Whole plan
Window
hour
Resets
Fixed UTC windows
On overage
On 429/503, wait at least the Retry-After time
Rollover
Not researched

The hourly cap for Pro. For the first 24 hours it is 200.

Official source for this limit
Requests, standard account20,000 requests / day
Scope
Whole plan
Window
day
Resets
Fixed UTC windows
On overage
On 429/503, wait at least the Retry-After time
Rollover
Not researched

The daily cap for Pro. For the first 24 hours it is 500.

Official source for this limit
Concurrent requests, standard account24 concurrent / concurrent
Scope
Whole plan
Window
concurrent
Resets
The slot frees up when the request finishes
On overage
429 after waiting for a free slot and timing out
Rollover
Not applicable

Concurrency for Pro. For the first 24 hours it is 2.

Official source for this limit

Officially confirmed · Checked

Official plan page for the latest terms
Official sources read (3)

Sources for this service