Compare/Nova Lite vs Granite 3.3 8B (Non-reasoning)

Nova LitevsGranite 3.3 8B (Non-reasoning)

Side-by-side comparison of pricing, 12 benchmarks, and generation speed.

Amazon

Nova Lite

Input
$0.06/M
Output
$0.24/M
Speed
190 tok/s
TTFT
0.39s
IBM

Granite 3.3 8B (Non-reasoning)

Input
$0.03/M
Output
$0.25/M
Speed
386 tok/s
TTFT
7.20s

Winner by Category

Cheaper
Granite 3.3 8B (Non-reasoning)
Faster (tok/s)
Granite 3.3 8B (Non-reasoning)
Lower Latency
Nova Lite
Benchmarks (12-0)
Nova Lite

Pricing Comparison

MetricNova LiteGranite 3.3 8B (Non-reasoning)
Input ($/M tokens)$0.06$0.03
Output ($/M tokens)$0.24$0.25
Cost for 1M input + 100K output tokens:
Nova Lite$0.08
Granite 3.3 8B (Non-reasoning)$0.06

Speed Comparison

Output Speed (tokens/s) — higher is better
Nova Lite
190 tok/s
Granite 3.3 8B (Non-reasoning)
386 tok/s
Time to First Token (seconds) — lower is better
Nova Lite
0.39s
Granite 3.3 8B (Non-reasoning)
7.20s

Benchmark Comparison

Data from Artificial Analysis API — 12 benchmarks

Intelligence Index
12.77.0
Coding Index
5.13.4
Math Index
7.06.7
GPQA Diamond
43.3%33.8%
MMLU-Pro
59.0%46.8%
LiveCodeBench
16.7%12.7%
AIME 2025
7.0%6.7%
MATH-500
76.5%66.5%
Humanity's Last Exam
4.6%4.2%
SciCode
13.9%10.1%
IFBench
34.1%22.4%
TerminalBench
0.8%0.0%
Nova Lite12 wins
0 winsGranite 3.3 8B (Non-reasoning)

Frequently Asked Questions

Which is cheaper, Nova Lite or Granite 3.3 8B (Non-reasoning)?

Granite 3.3 8B (Non-reasoning) is cheaper overall. Its blended price (3:1 input/output ratio) is $0.09/M tokens vs $0.10/M for Nova Lite.

Which model performs better on benchmarks?

Nova Lite wins 12 out of 12 benchmarks compared to 0 for Granite 3.3 8B (Non-reasoning). See the detailed benchmark chart above for per-category results.

Which is faster for real-time applications?

Granite 3.3 8B (Non-reasoning) generates tokens faster at 386 tok/s vs 190 tok/s. Nova Lite also has lower time-to-first-token (0.39s vs 7.20s).

When should I use Nova Lite vs Granite 3.3 8B (Non-reasoning)?

Choose based on your priorities: Granite 3.3 8B (Non-reasoning) for lower cost, Nova Lite for stronger benchmark performance, and Granite 3.3 8B (Non-reasoning) for faster generation. For latency-sensitive apps, check the TTFT comparison above.