Compare/Mercury 2 vs Llama 3.3 Instruct 70B

Mercury 2vsLlama 3.3 Instruct 70B

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

Inception

Mercury 2

Input
$0.25/M
Output
$0.75/M
Speed
1108 tok/s
TTFT
3.08s
Meta

Llama 3.3 Instruct 70B

Input
$0.58/M
Output
$0.71/M
Speed
88 tok/s
TTFT
0.65s

Winner by Category

Cheaper
Mercury 2
Faster (tok/s)
Mercury 2
Lower Latency
Llama 3.3 Instruct 70B
Benchmarks (6-6)
Tie

Pricing Comparison

MetricMercury 2Llama 3.3 Instruct 70B
Input ($/M tokens)$0.25$0.58
Output ($/M tokens)$0.75$0.71
Cost for 1M input + 100K output tokens:
Mercury 2$0.33
Llama 3.3 Instruct 70B$0.65

Speed Comparison

Output Speed (tokens/s) — higher is better
Mercury 2
1108 tok/s
Llama 3.3 Instruct 70B
88 tok/s
Time to First Token (seconds) — lower is better
Mercury 2
3.08s
Llama 3.3 Instruct 70B
0.65s

Benchmark Comparison

Data from Artificial Analysis API — 12 benchmarks

Intelligence Index
25.39.4
Coding Index
11.9
Math Index
7.7
GPQA Diamond
77.0%49.8%
MMLU-Pro
71.3%
LiveCodeBench
28.8%
AIME 2025
7.7%
MATH-500
77.3%
Humanity's Last Exam
15.5%4.0%
SciCode
38.7%26.0%
IFBench
69.8%47.1%
TerminalBench
26.5%3.0%
Mercury 26 wins
6 winsLlama 3.3 Instruct 70B

Frequently Asked Questions

Which is cheaper, Mercury 2 or Llama 3.3 Instruct 70B?

Mercury 2 is cheaper overall. Its blended price (3:1 input/output ratio) is $0.38/M tokens vs $0.61/M for Llama 3.3 Instruct 70B.

Which model performs better on benchmarks?

It's a tie — both models win 6 benchmarks each across 12 evaluated categories. See the detailed benchmark chart above for per-category results.

Which is faster for real-time applications?

Mercury 2 generates tokens faster at 1108 tok/s vs 88 tok/s. However, Llama 3.3 Instruct 70B has lower time-to-first-token (0.65s vs 3.08s).

When should I use Mercury 2 vs Llama 3.3 Instruct 70B?

Choose based on your priorities: Mercury 2 for lower cost, both perform similarly on benchmarks, and Mercury 2 for faster generation. For latency-sensitive apps, check the TTFT comparison above.