Model comparison

Codestral vs Mistral Medium 3.5

Mistral Medium 3.5 is the stronger model overall, scoring 40.2 to 30.6 on the Noometry Index. Codestral costs 6.7× less per token, which makes it the better buy when Mistral Medium 3.5's lead doesn't matter for your workload.

Last verified October 10, 2026. 1 shared benchmarks.

Codestral

Mistral AI

30.6

Rank #290 Reported

Mistral Medium 3.5

Mistral AI

40.2

Rank #152 Confirmed

Summary

They share 1 benchmark with published results for both. Codestral scores higher in 1 category and Mistral Medium 3.5 in 1 category; 2 gaps are clear of the uncertainty.
The widest gap is in coding, where Mistral Medium 3.5 leads 36.0 to 27.3.
The biggest single-benchmark swing is Kagi LLM Benchmark: 32.5% for Codestral and 41.4% for Mistral Medium 3.5.
Codestral is cheaper at $0.30 / $0.90 per million input/output tokens, against $1.50 / $7.50 for Mistral Medium 3.5.
Mistral Medium 3.5 accepts more context: 262K tokens versus 256K.
Mistral Medium 3.5 has downloadable open weights; the other is API-only.

Side by side

Codestral and Mistral Medium 3.5 specifications
	Codestral	Mistral Medium 3.5
Provider	Mistral AI	Mistral AI
Noometry Index	30.6	40.2
Released	2024-05-29	—
Weights	Proprietary	Open
Context window	256K	262K
Max output	8K	210K
Input $ / M tokens	$0.30	$1.50
Output $ / M tokens	$0.90	$7.50
Results tracked	7	22

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

Category by category

Coding Mistral Medium 3.5 leads

Codestral: 27.3 (#321), Mistral Medium 3.5: 36.0 (#213)

Coding benchmarks
Benchmark	Codestral	Mistral Medium 3.5
Aider Polyglot	11.1%	—
LMArena WebDev	—	1264
BigCodeBench Instruct	41.8%	—
LMArena Coding	—	1461
BigCodeBench Complete	52.5%	—
ALE-Bench	137.78	—
HumanEval+	73.8%	—
MBPP+	61.9%	—

Reasoning Codestral leads

Codestral: 19.8 (#251), Mistral Medium 3.5: 17.3 (#295)

Reasoning benchmarks
Benchmark	Codestral	Mistral Medium 3.5
Kagi LLM Benchmark	32.5%	41.4%
NYT Connections (extended)	—	12.9%
LMArena Hard Prompts	—	1436
Epoch Capabilities Index	—	141.35

Math Not comparable

Codestral: —, Mistral Medium 3.5: 39.1 (#113)

Math benchmarks
Benchmark	Codestral	Mistral Medium 3.5
LMArena Math	—	1431

Knowledge Not comparable

Codestral: —, Mistral Medium 3.5: 40.0 (#126)

Knowledge benchmarks
Benchmark	Codestral	Mistral Medium 3.5
LMArena Expert	—	1432

Multimodal Not comparable

Codestral: —, Mistral Medium 3.5: 38.3 (#65)

Multimodal benchmarks
Benchmark	Codestral	Mistral Medium 3.5
LMArena Vision	—	1223

Multilingual Not comparable

Codestral: —, Mistral Medium 3.5: 51.9 (#100)

Multilingual benchmarks
Benchmark	Codestral	Mistral Medium 3.5
LMArena Non-English	—	1404
LMArena Chinese	—	1442
LMArena French	—	1448
LMArena German	—	1451
LMArena Korean	—	1385
LMArena Russian	—	1395
LMArena Spanish	—	1409

Instruction Following Not comparable

Codestral: —, Mistral Medium 3.5: 74.6 (#90)

Instruction Following benchmarks
Benchmark	Codestral	Mistral Medium 3.5
LMArena Instruction Following	—	1415

Long Context Not comparable

Codestral: —, Mistral Medium 3.5: 43.2 (#103)

Long Context benchmarks
Benchmark	Codestral	Mistral Medium 3.5
LMArena Longer Query	—	1415

Writing & Preference Not comparable

Codestral: —, Mistral Medium 3.5: 58.5 (#117)

Writing & Preference benchmarks
Benchmark	Codestral	Mistral Medium 3.5
LMArena Text	—	1421
LMArena Creative Writing	—	1374
EQ-Bench 4	—	993
LMArena Multi-Turn	—	1423

Frequently asked questions

Is Codestral better than Mistral Medium 3.5?

Which is cheaper, Codestral or Mistral Medium 3.5?

Codestral is cheaper. It lists at $0.30 per million input tokens and $0.90 per million output tokens; Mistral Medium 3.5 lists at $1.50 and $7.50.

Is Codestral or Mistral Medium 3.5 better for coding?

Mistral Medium 3.5 scores higher on coding benchmarks: 36.0 versus 27.3 in the Noometry coding category.

Which has the bigger context window?

Mistral Medium 3.5 does, with 262K tokens against 256K.

How many benchmarks do Codestral and Mistral Medium 3.5 share?

1 benchmark has published results for both models. Codestral has 7 scored results on Noometry and Mistral Medium 3.5 has 22.