Inference
EN
Mistral

Mistral Medium 3.5 128B

Mistral Medium 3.5 128B is a reasoning model from Mistral in the Mistral Medium 3.5 family. 13 benchmarks count toward its index, in 6 categories.

  • Open weights
  • Reasoning
  • Input: text
  • Released 2026-04
RankedEnough results for an index.
IndexAll categories, out of 100.50.3±6.0Median 51.0
CoverageIndex weight that has results.85%
SpeedOutput tokens per second.—Median 67/s
First tokenSeconds to the first output token.—Median 1.3 s
Input / 1MUS dollars per 1M input tokens.$1.50Median $0.60
Output / 1MUS dollars per 1M output tokens.$7.50Median $2.55
ContextMaximum tokens in one request.256KMedian 262K
EloLMArena rating and rank.—Median 1422

Run locally

127.7B parameters
16-bit
269.2 GB
8-bit
143.5 GB
4-bit
82.3 GB
Weights
Hugging Face (opens in a new tab) ↗

Memory need with a short context. The 4-bit copy fits in 128 GB. Compare local models

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.4 counted60.166 / 218
CodingCode writing and repair.2 counted53.4112 / 209
ReasoningLogic problems and puzzles.2 counted45.8140 / 227
MultimodalTasks with images and text.1 counted39.6—
KnowledgeFacts and expert knowledge.3 counted41.2187 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted53.3—
MathMath problems.None counted——

Results

13 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
τ³-Bench Tool-Agent-User EvaluationAgenticAbove fitFar above this model's index.91.4%2 / 32
τ²-Bench Tool-Agent-User EvaluationAgentic94.2%26 / 149
Software Engineering Benchmark VerifiedCoding77.6%33 / 114
Artificial Analysis IFBenchInstruction68.8%49 / 146
Artificial Analysis Long Context ReasoningReasoning69.3%115 / 217
Artificial Analysis Humanity's Last ExamKnowledge13.8%130 / 222
Artificial Analysis GPQA DiamondKnowledge74.8%119 / 199
GDPval-AA normalizedAgentic13.0%91 / 133
Critical Physics TasksReasoning0.0%150 / 214
Artificial Analysis Omniscience AccuracyKnowledge23.0%116 / 164
Artificial Analysis MMMU-ProMultimodal64.9%84 / 113
Artificial Analysis SciCodeCoding40.2%87 / 114
Gert Labs Composite Game BenchmarkAgentic39.1%40 / 52

Sources

BenchLMCC BY-NC 4.0

History

5 releases
Index51.0 → 50.3
23 Oct 202612 Nov 2026
Output / 1M$7.50 → $7.50
23 Oct 202612 Nov 2026

Questions

How much does Mistral Medium 3.5 128B cost?
$1.50 per 1M input tokens, $7.50 per 1M output tokens.
How good is Mistral Medium 3.5 128B?
Its index is 50.3 out of 100. That is place 148 of 283 ranked models.
What is Mistral Medium 3.5 128B best at?
Agentic: 60.1 out of 100, place 66 of 218.
What is the context window of Mistral Medium 3.5 128B?
256K tokens.
Is Mistral Medium 3.5 128B open source?
Yes. Its weights are published.
When was Mistral Medium 3.5 128B released?
April 2026.

Badge

Mistral Medium 3.5 128B on Inference360: place 148 of 283, index 50.3
HTML
<a href="https://inference360.ai/models/mistral-medium-3-5-128b"><img src="https://inference360.ai/badges/mistral-medium-3-5-128b.svg" alt="Mistral Medium 3.5 128B on Inference360: place 148 of 283, index 50.3" height="20"></a>
Markdown
[![Mistral Medium 3.5 128B on Inference360: place 148 of 283, index 50.3](https://inference360.ai/badges/mistral-medium-3-5-128b.svg)](https://inference360.ai/models/mistral-medium-3-5-128b)

Compare with

GLM-4.750.0Qwen3.6-35B-A3B50.6Nemotron 3 Ultra50.8Mistral Small 437.9

Lower price

dots3-note Preview63.4 · FreeApodex 1.1 Mini55.7 · FreeNemotron 3 Ultra50.8 · Free

More from Mistral

All models
Mistral Medium 3.542.7Mistral Small 437.9LeanstralUnrankedMistral Small 4 (Reasoning)UnrankedLeanstral 1.5Unranked