Model comparison

Codestral vs Nemotron 3 Ultra

Nemotron 3 Ultra is the stronger model overall, scoring 44.7 to 30.1 on the Noometry Index. Codestral costs 2.3× less per token, which makes it the better buy when Nemotron 3 Ultra's lead doesn't matter for your workload.

Last verified . 0 shared benchmarks.

Codestral Mistral AI

30.1

Rank #287 Reported

Nemotron 3 Ultra NVIDIA

44.7

Rank #71 Confirmed

Summary

  • The widest gap is in reasoning, where Nemotron 3 Ultra leads 30.3 to 19.1.
  • Codestral is cheaper at $0.30 / $0.90 per million input/output tokens, against $0.60 / $2.40 for Nemotron 3 Ultra.
  • Nemotron 3 Ultra accepts more context: 262K tokens versus 256K.

Side by side

Codestral and Nemotron 3 Ultra specifications
CodestralNemotron 3 Ultra
ProviderMistral AINVIDIA
Noometry Index30.144.7
Released2024-05-292026-06-04
WeightsOpenOpen
Context window256K262K
Max output4K128K
Input $ / M tokens$0.30$0.60
Output $ / M tokens$0.90$2.40
Results tracked729

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding Nemotron 3 Ultra leads

Codestral: 26.9 (#316), Nemotron 3 Ultra: 37.7 (#174)

Coding benchmarks
BenchmarkCodestralNemotron 3 Ultra
FrontierCode—13.6%
Aider Polyglot11.1%—
SciCode—40.3%
WeirdML—43.5%
BigCodeBench Instruct41.8%—
LMArena Coding—1468
BigCodeBench Complete52.5%—
ALE-Bench137.78—
HumanEval+73.8%—
MBPP+61.9%—

Reasoning Nemotron 3 Ultra leads

Codestral: 19.1 (#244), Nemotron 3 Ultra: 30.3 (#78)

Reasoning benchmarks
BenchmarkCodestralNemotron 3 Ultra
Kagi LLM Benchmark32.5%—
CritPt—3.1%
Chess Puzzles—12%
LMArena Hard Prompts—1452
Mystery Game Puzzles—20%
DTBench—90.1%
LMCA—36.9%
Epoch Capabilities Index—146.2

Math Not comparable

Codestral: —, Nemotron 3 Ultra: 34.4 (#160)

Math benchmarks
BenchmarkCodestralNemotron 3 Ultra
OTIS Mock AIME 2024-2025—86.7%
ProofBench—2%
LMArena Math—1456

Knowledge Not comparable

Codestral: —, Nemotron 3 Ultra: 52.5 (#56)

Knowledge benchmarks
BenchmarkCodestralNemotron 3 Ultra
GPQA Diamond—85.4%
LMArena Expert—1472

Multilingual Not comparable

Codestral: —, Nemotron 3 Ultra: 53.7 (#61)

Multilingual benchmarks
BenchmarkCodestralNemotron 3 Ultra
LMArena Non-English—1427
LMArena Chinese—1498
LMArena French—1475
LMArena German—1471
LMArena Korean—1387
LMArena Russian—1415
LMArena Spanish—1452

Instruction Following Not comparable

Codestral: —, Nemotron 3 Ultra: 74.9 (#84)

Instruction Following benchmarks
BenchmarkCodestralNemotron 3 Ultra
LMArena Instruction Following—1417

Long Context Not comparable

Codestral: —, Nemotron 3 Ultra: 43.9 (#77)

Long Context benchmarks
BenchmarkCodestralNemotron 3 Ultra
LMArena Longer Query—1437

Writing & Preference Not comparable

Codestral: —, Nemotron 3 Ultra: 66.3 (#35)

Writing & Preference benchmarks
BenchmarkCodestralNemotron 3 Ultra
LMArena Text—1444
LMArena Creative Writing—1392
EQ-Bench Creative Writing—1692
LMArena Multi-Turn—1412

Frequently asked questions

Is Codestral better than Nemotron 3 Ultra?

Nemotron 3 Ultra is the stronger model overall, scoring 44.7 to 30.1 on the Noometry Index. Codestral costs 2.3× less per token, which makes it the better buy when Nemotron 3 Ultra's lead doesn't matter for your workload.

Which is cheaper, Codestral or Nemotron 3 Ultra?

Codestral is cheaper. It lists at $0.30 per million input tokens and $0.90 per million output tokens; Nemotron 3 Ultra lists at $0.60 and $2.40.

Is Codestral or Nemotron 3 Ultra better for coding?

Nemotron 3 Ultra scores higher on coding benchmarks: 37.7 versus 26.9 in the Noometry coding category.

Which has the bigger context window?

Nemotron 3 Ultra does, with 262K tokens against 256K.

How many benchmarks do Codestral and Nemotron 3 Ultra share?

0 benchmarks have published results for both models. Codestral has 7 scored results on Noometry and Nemotron 3 Ultra has 29.

Related comparisons

Go deeper