Granite 4.0 Micro
Granite 4.0 MicroRankingin non-thinking mode?
Alternatives?
CheaperGPT OSS 20BlowOpen+95%-12%$0.018 / $0.09Memory (4-bit): 1.6 GB → 13 GBBetterGPT-6 Lunalow×4.2+391%$0.005Function calling only with none · Chat Completions+3IOpenK2 Horizon 7Bthinking×4+1,188%$0.30 / $1.20Memory (4-bit): 1.6 GB → 4.5 GB
CheaperQwen3.5 0.8BthinkingOpen+20%-51%$0.01 / $0.05BetterGranite 4.2 3BthinkingOpen+77%+29%$0.006Context: 131K → 128K+1GPT OSS 20BlowOpen+95%-12%$0.018 / $0.09Memory (4-bit): 1.6 GB → 13 GBQwen3.5 0.8BthinkingOpen+20%-51%$0.01 / $0.05GPT-6 Lunalow×4.2+391%$0.005Function calling only with none · Chat Completions+3Granite 4.2 3BthinkingOpen+77%+29%$0.006Context: 131K → 128K+1
Gemma 4 E4BthinkingOpen+74%same$0.008Context: 131K → 128K+1GPT-5.6 Lunalow×4.1+1,004%$0.01GPT OSS 20BhighOpen+76%-12%$0.011Memory (4-bit): 1.6 GB → 13 GBGemma 4 12BthinkingOpen×2.8+268%$0.10 / $0.30Memory (4-bit): 1.6 GB → 6 GBQwen 3.5 9Bnon-thinkingOpen×2.6+127%$0.08 / $0.13Memory (4-bit): 1.6 GB → 4.8 GBQwen3.5 4BthinkingOpen×2.6+47%$0.03 / $0.15Memory (4-bit): 1.6 GB → 2.3 GBGranite 4.2 30BthinkingOpen×2.5—$0.072Context: 131K → 128K+1Granite 4.2 8BthinkingOpen×2.2—$0.024Context: 131K → 128K+1Granite 4.2 3BthinkingOpen+77%—$0.006Context: 131K → 128K+1Granite 4.1 30BOpen+46%—$0.10 / $0.25Memory (4-bit): 1.6 GB → 17 GBGranite 4.1 8Bnon-thinkingOpen+29%——Memory (4-bit): 1.6 GB → 4.4 GBGranite 4.0 H Smallnon-thinkingOpensame—$0.05 / $0.22Context: 131K → 128K+1Granite 4.1 3BOpensame——Memory (4-bit): 1.6 GB → 1.7 GBGranite 4.0 H 1BOpensame——Context: 131K → 128KAll IBM models →Run it yourself
estimateBenchmarks
all categoriesFAQ
How much does Granite 4.0 Micro cost?›
$0.017 per 1M input tokens and $0.11 per 1M output tokens (cheapest host: Cloudflare).
Is Granite 4.0 Micro free to use?›
There is no free API; the cheapest paid price is $0.017 per 1M input tokens. The weights are open, so you can run it yourself.
How good is Granite 4.0 Micro, and what is it best at?›
#272 of 334 in Chat & Knowledge, #195 of 239 in Coding, #216 of 242 in Agents & Tool Use, #230 of 245 in Writing & Language.
Is Granite 4.0 Micro better than Apertus 70B Instruct?›
On AA Intelligence Index 4.3.2, Granite 4.0 Micro scores 5.1 vs 5.1 — about the same.
How can I use Granite 4.0 Micro?›
Through the API of Cloudflare.
What is the context window of Granite 4.0 Micro?›
131K tokens, with up to 118K output tokens.
When was Granite 4.0 Micro released? Is there a newer version?›
Released 2025-09-22. It is the current version in its line.
Is Granite 4.0 Micro open source?›
Yes, the weights are open (Apache-2.0 licence).
Can I run Granite 4.0 Micro locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can Granite 4.0 Micro read images?›
No, it accepts text only.
What is Granite 4.0 Micro?›
Granite 4.0 Micro is an open-weights language model from IBM, released Sep 2025.
What are the best alternatives to Granite 4.0 Micro?›
GPT OSS 20B (cheaper), GPT-6 Luna (better), K2 Horizon 7B (open weights).