IB
IBM
Language model · United States ·Apr 2026

IBM Granite 4.1

Language Open Source
30B
Parameters
512K
Context
7
Benchmarks

About IBM Granite 4.1

IBM Granite 4.1 is an AI model developed by IBM, in the language category, released in 2026, made available as an open-source model with 30B params. Granite 4.1 is IBM's latest open family of dense language models in 3B, 8B, and 30B sizes (base and instruct), released .

On this page you'll find IBM Granite 4.1's full specifications, including a context window of 512K tokens, 7 benchmark results. Review provider pricing and benchmark scores below, or compare IBM Granite 4.1 head-to-head with other language models.

→ More models by IBM → All language models → AI benchmarks leaderboard

Run IBM Granite 4.1 locally — what to buy

Share this build ↗

Sized to what IBM Granite 4.1 actually needs (~512k context at Q4), not to the biggest GPU: Good = cheapest that runs it, Better = best value, Best = most headroom. Speed is a hardware estimate; anything past ~120 tok/s is shown as instant.

Good
AMD Radeon RX 7900 XTX
Consumer GPU · ~$999
~44 tok/s
Q4 · fits full context · up to 14k ctx
Runs with: ExLlamaV2, KoboldCpp, LM Studio
See full estimate →
Better
NVIDIA GeForce RTX 4090
Consumer GPU · ~$1,599
~52.2 tok/s
Q4 · fits full context · up to 14k ctx
Runs with: ExLlamaV2, KoboldCpp, LM Studio
See full estimate →
Best
NVIDIA GeForce RTX 5090
Consumer GPU · ~$1,999
~92.8 tok/s
Q4 · fits full context · up to 42k ctx
Runs with: ExLlamaV2, KoboldCpp, LM Studio
See full estimate →

Software support

·ExLlamaV2 ·KoboldCpp ·LM Studio ·MLX ·Ollama ·SGLang ·TensorRT-LLM ·llama.cpp ·vLLM ?KTransformers ?text-generation-webui

✓ measured · · compatible · — not supported. Informational only — speed is hardware-based.

Consumer GPU
11.7–92.8 tok/s
9 options →
Apple
14.6–39 tok/s
17 options →
Datacenter GPU
39.6–120+ tok/s
13 options →

Community

Would you run IBM Granite 4.1 again?
No votes yet — be the first.
Did it run? — community reports by hardware

No reports yet. Own this model on some hardware? Be the first to confirm it runs.

Discussion

No comments yet. Start the discussion.

IBM Granite 4.1 benchmark scores▶
Compare on the benchmark table →
Benchmark Category Score Variant
HumanEval code 85.37%
MBPP code 87.3%
IFEval general 87.06%
MMLU general 73.84%
MMLU-Pro general 55.99%
SimpleQA safety 4.82%
BFCL v3 tool-use 68.27%

More models like IBM Granite 4.1

AI
DMind-1
 AI
COLLECTIVE-1
 AI
Cogito 1 series
 AI
Euler (Euler大模型)
· 2.5B
 AI
SmolLM-1.7B
· 1.7B
 AI
Llama-3-70B-Synthia-v3.5
· 70B

Frequently asked questions

What is IBM Granite 4.1? ▶

Granite 4.1 is IBM's latest open family of dense language models in 3B, 8B, and 30B sizes (base and instruct), released It is tracked on GenAIList with its specifications, benchmark scores and provider pricing.

Who created IBM Granite 4.1? ▶

IBM Granite 4.1 was developed by IBM and released in 2026.

Is IBM Granite 4.1 open source or proprietary? ▶

IBM Granite 4.1 is released as open source — its weights are publicly available to download, run locally and fine-tune.

What is IBM Granite 4.1's context window? ▶

IBM Granite 4.1 supports a context window of 512K tokens, and has 30B params.

How much does IBM Granite 4.1 cost? ▶

Pricing for IBM Granite 4.1 depends on the provider. See the providers table on this page for the latest API rates.

How does IBM Granite 4.1 perform on benchmarks? ▶

IBM Granite 4.1 is benchmarked across 7 evaluations on GenAIList, including HumanEval (85.37%). See the full benchmark table below and compare it with other models.

IBM Granite 4.1 8B - Open Source AI That Actually Surprised Me
Granite 4.1, IBM Bob & building a quantum ecosystem
Granite 4.1 - The Fastest ASR?

Know a great video about IBM Granite 4.1?