Inferse
For vendors
AdvertiseTop spots, featured rows, home sponsor Get listedAdd a product we don't cover yet Claim a listingFree: verify the domain, then correct your page Vendor dashboardNumbers, orders and placements
List your platform
Vendor dashboard

Granite 4.0 Micro

IBM· released 20 Oct 2025

ibm-granite/granite-4.0-h-micro
Input / 1M tokens
$0.017
Output / 1M tokens
$0.112
Context window131K tokens

Longest answer: 118K tokens

reads textopen weights
See it on openrouter.ai

About

Granite 4.0 Micro is IBM’s 3B-parameter text model in the ibm-granite family. Released on 2025-10-20, it has open weights and accepts text inputs to produce text outputs. Its context window is 131,000 tokens, and its maximum output is 117,900 tokens. The model is identified as ibm-granite/granite-4.0-h-micro on both the Hugging Face listing and model ID. IBM is the maker and lab. Listed API pricing is $0.017 USD per 1M input tokens and $0.112 USD per 1M output tokens. The model’s long context and text-only interface define its listed scope: the available facts specify no image, audio, or other modality. Builders can assess its per-token input and output costs separately when estimating text workloads, alongside the output limit and context size.

Who it is for

It suits teams evaluating an open-weight IBM model for text input and output, particularly where a 131,000-token context window is relevant.

What is good

  • Open weights are available.
  • 131,000-token context window.
  • 117,900-token maximum output.
  • Separate input and output token pricing.

What to know first

  • Text-only input and output.
  • Output pricing is $0.112 USD per 1M tokens.

Verdict

Granite 4.0 Micro offers open weights and a large listed context window for text workloads. Account for its distinct input and output rates and maximum output when estimating usage.

Details

Lab
IBMopenrouter.ai · 3 Oct 2026
Context
131Kopenrouter.ai · 3 Oct 2026
Input price
$0.017 / 1Mopenrouter.ai · 3 Oct 2026
Output price
$0.112 / 1Mopenrouter.ai · 3 Oct 2026
Max output
117,900 tokensopenrouter.ai · 3 Oct 2026
Inputs
textopenrouter.ai · 3 Oct 2026
Open weights
Yesopenrouter.ai · 3 Oct 2026

More longest-context models

See the list
Auto Router ctx2,000KinVaries by model routed tooutVaries by model routed to openrouterAuto Router (Beta) ctx2,000KinVaries by model routed tooutVaries by model routed to openrouterGrok 4.20 ctx2,000Kin$1.25 / 1Mout$2.5 / 1M SpaceXAIGrok 4.20 Multi-Agent ctx2,000Kin$1.25 / 1Mout$2.5 / 1M SpaceXAIPareto Code Router ctx2,000KinVaries by model routed tooutVaries by model routed to openrouterLlama 4 Scout ctx1,311Kin$0.1 / 1Mout$0.3 / 1M Meta

Listed on Inferse

Sources