Skip to main content
Back to Explore

Granite 3.3 8B (Non-reasoning)

IBM·Released 2025-04-16
Open Source

Pricing

Input

$0.03

per 1M tokens

Output

$0.25

per 1M tokens

Blended

$0.09

per 1M tokens

Cheaper than 86% of models. Median price is $0.54/1M tokens.

Cost Calculator

Tokens per day1M
100K100M

Daily

$0.09

Monthly

$2.55

vs. Similar Models

LFM2 8B A1BQ:0.0
$0.01-85%
Command-R (Mar '24)Q:+0.3
$0.75+782%
Mistral 7B InstructQ:+0.3
$0.21+142%
Qwen3 1.7B (Non-reasoning)Q:-0.3
$0.19+121%

Performance

434

tokens/sec

Faster than 98% of models

20.40

seconds

Faster than 5% of models

20.40

seconds

Faster than 24% of models

Market Median

95 tok/s

359% faster

Median TTFT

1.11s

1743% slower

Throughput/Dollar

5110

tok/s per $/1M

Speed Comparison

HyperNova 60B 2605
417 tok/s-4%
StepFun: Step 3.7 Flash
414 tok/s-5%
Granite 4.0 H Small
481 tok/s+11%

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

Compare all 7 models