Granite 3.3 8B (Non-reasoning)

IBM·Released 2025-04-16

Open Source

Model Family

Granite 4.0 H 1B2025-10-28 Granite 4.0 1B2025-10-28 Granite 4.0 350M2025-10-28 Granite 4.0 H 350M2025-10-28 IBM: Granite 4.0 Micro2025-10-20 Granite 4.0 H Small2025-09-22 Granite 4.0 Micro2025-09-22

Quality Index

7.0

426th of 442

Top 97%

Coding Index

3.4

319th of 352

Top 91%

Math Index

6.7

240th of 268

Top 90%

Price/1M

$0.09

215th cheapest

73% below median

Top 32%

Speed

443 tok/s

Top 1%

TTFT

7.32s

Market Position

Granite 3.3 8B (Non-reasoning)Market Average

Pricing

Input

$0.03

per 1M tokens

Output

$0.25

per 1M tokens

Blended

$0.09

per 1M tokens

Cheaper than 68% of models. Median price is $0.31/1M tokens.

Cost Calculator

Tokens per day1M

100K100M

Daily

$0.09

Monthly

$2.55

vs. Similar Models

LFM2 8B A1BQ:0.0

$0.00-100%

Qwen3 1.7B (Non-reasoning)Q:-0.2

$0.19+121%

Molmo2-8BQ:+0.3

$0.00-100%

Granite 4.0 1BQ:+0.3

$0.00-100%

Performance

443

tokens/sec

Faster than 99% of models

7.32

seconds

Faster than 8% of models

7.32

seconds

Faster than 23% of models

Market Median

46 tok/s

871% faster

Median TTFT

0.42s

1651% slower

Throughput/Dollar

5209

tok/s per $/1M

Speed Comparison

NVIDIA Nemotron 3 Super 120B A12B (Reasoning)

408 tok/s-8%

Granite 4.0 H Small

496 tok/s+12%

Gemini 2.5 Flash-Lite Preview (Sep '25) (Reasoning)

356 tok/s-20%

Benchmarks

MMLU-Pro

46.8%

GPQA Diamond

33.8%

HLE

4.2%

LiveCodeBench

12.7%

SciCode

10.1%

TerminalBench Hard

0.0%

MATH-500

66.5%

AIME

4.7%

AIME 2025

6.7%

IFBench

22.4%

Long Context Recall

4.3%

Tau2

10.5%

Market AverageTop Score

Open Source

Quick Compare

Similar Models

LFM2 8B A1B

Liquid AI

Q: 7.0$0.00/1M33K ctx

Slower: 100%Cheaper: 100%

Qwen3 1.7B (Non-reasoning)

Alibaba

Q: 6.8$0.19/1M

Slower: 69%Pricier: 121%

Molmo2-8B

Allen AI

Q: 7.3$0.00/1M

Slower: 69%Cheaper: 100%

Granite 4.0 1B

IBM

Q: 7.3$0.00/1M

Slower: 100%Cheaper: 100%

Llama 65B