granite-3.1-1b-a400m-instruct

A Core AI bundle of ibm-granite/granite-3.1-1b-a400m-instruct for the Aether SDK (iOS and macOS 27+).

• Source: ibm-granite/granite-3.1-1b-a400m-instruct at revision 0da7a48b0276d500ce5922fd2b33944091fc6c09, licence Apache-2.0. The licence is included as LICENSE. ibm-granite/granite-3.1-1b-a400m-instruct declares apache-2.0 in its model card but ships no licence file; LICENSE is the canonical text from apache.org.
• Changes from the source: converted from PyTorch to Core AI (.aimodel) by Aether forge (recipe granite-3.1-1b-a400m-instruct@2). Weights are int8-linear-perblock16 (8-bit weights). The tokenizer files are the source's own.

Variants

Variant	Platform	Arch	Compute	Compiled	Assets	Download
macos-any-gpu	macos	any	gpu	no (specialized on first load)	granite_3_1_1b_a400m_instruct.aimodel 1.5 GB	1.51 GB
ios-any-gpu	ios	any	gpu	no (specialized on first load)	granite_3_1_1b_a400m_instruct.aimodel 1.5 GB	1.51 GB

Verification

Every row is a record in verification/ about exactly these bytes (matched by bundle digest). Reference rows are strict T2 passes of the unquantized export on the same fixture, in verification/reference/.

Variant	Tier	Result	Detail	Device	OS build	Compute	Record
ios-any-gpu	T0	pass		iPhone18,2	24A446	target	e293a5ad
ios-any-gpu	T2	pass	18/18 strict; profile quantized-8bit; fixture 18faeb8faf71f791	iPhone18,2	24A446	target	3faf5f4b
ios-any-gpu	T3	pass	copy-fidelity-v1; 82.0% vs reference 80.0%; 50 items	iPhone18,2	24A446	target	2eb79dd3
macos-any-gpu	T0	pass		Mac17,6	26A434	target	87e18d09
macos-any-gpu	T1	pass		Mac17,6	26A434	target	e64f7051
macos-any-gpu	T2	pass	18/18 strict; 18/18 strict; profile quantized-8bit; fixture 18faeb8faf71f791	Mac17,6	26A434	target	e812cb45
macos-any-gpu	T3	pass	gsm8k-test-500; 11.2% vs reference 11.4%; 500 items	Mac17,6	26A434	target	957a53b6
macos-any-gpu	T3	pass	copy-fidelity-v1; 82.0% vs reference 80.0%; 50 items	Mac17,6	26A434	target	973a7e26
unquantized reference (not published)	T2	pass	18/18 strict; 18/18 strict; profile strict; fixture 18faeb8faf71f791	Mac17,6	26A434	target	86cf7e26

The quantized profile also requires: T2 strict on the unquantized reference export (met by the reference row).

Use

aether run granite-3.1-1b-a400m-instruct --prompt "Hello"

import Aether

let aether = try Aether()
let chat = try await aether.chat("granite-3.1-1b-a400m-instruct")
let reply = try await chat.respond(to: "Hello")
print(reply.text)