Compare/Devstral Small (May '25) vs Nova Micro

Devstral Small (May '25)vsNova Micro

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

Mistral

Devstral Small (May '25)

Input
$0.06/M
Output
$0.12/M
Speed
TTFT
Amazon

Nova Micro

Input
$0.035/M
Output
$0.14/M
Speed
303 tok/s
TTFT
0.35s

Winner by Category

Cheaper
Nova Micro
Faster (tok/s)
Nova Micro
Lower Latency
Devstral Small (May '25)
Benchmarks (8-4)
Devstral Small (May '25)

Pricing Comparison

MetricDevstral Small (May '25)Nova Micro
Input ($/M tokens)$0.06$0.035
Output ($/M tokens)$0.12$0.14
Cost for 1M input + 100K output tokens:
Devstral Small (May '25)$0.07
Nova Micro$0.05

Speed Comparison

Output Speed (tokens/s) — higher is better
Devstral Small (May '25)
Nova Micro
303 tok/s
Time to First Token (seconds) — lower is better
Devstral Small (May '25)
Nova Micro
0.35s

Benchmark Comparison

Data from Artificial Analysis API — 12 benchmarks

Intelligence Index
18.010.3
Coding Index
12.24.1
Math Index
6.0
GPQA Diamond
43.4%35.8%
MMLU-Pro
63.2%53.1%
LiveCodeBench
25.8%14.0%
AIME 2025
6.0%
MATH-500
68.4%70.3%
Humanity's Last Exam
4.0%4.7%
SciCode
24.5%9.4%
IFBench
31.6%29.4%
TerminalBench
6.1%1.5%
Devstral Small (May '25)8 wins
4 winsNova Micro

Frequently Asked Questions

Which is cheaper, Devstral Small (May '25) or Nova Micro?

Nova Micro is cheaper overall. Its blended price (3:1 input/output ratio) is $0.06/M tokens vs $0.07/M for Devstral Small (May '25).

Which model performs better on benchmarks?

Devstral Small (May '25) wins 8 out of 12 benchmarks compared to 4 for Nova Micro. See the detailed benchmark chart above for per-category results.

Which is faster for real-time applications?

Nova Micro generates tokens faster at 303 tok/s vs 0 tok/s. Devstral Small (May '25) also has lower time-to-first-token (0.00s vs 0.35s).

When should I use Devstral Small (May '25) vs Nova Micro?

Choose based on your priorities: Nova Micro for lower cost, Devstral Small (May '25) for stronger benchmark performance, and Nova Micro for faster generation. For latency-sensitive apps, check the TTFT comparison above.