Inference
EN
Mistral

Mistral Large 3

Mistral Large 3 is a standard model from Mistral. 17 benchmarks count toward its index, in 7 categories.

  • Open weights
  • Standard
  • Input: text · image · file
  • Released 2025-12
RankedEnough results for an index.
IndexAll categories, out of 100.34.3±4.4Median 51.0
CoverageIndex weight that has results.95%
SpeedOutput tokens per second.61/sMedian 67/s
First tokenSeconds to the first output token.0.7 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.$0.50Median $0.60
Output / 1MUS dollars per 1M output tokens.$1.50Median $2.55
ContextMaximum tokens in one request.262KMedian 262K
EloLMArena rating and rank.1429±3 #98Median 1422

Details

From OpenRouter
API IDThe model ID on OpenRouter.
mistralai/mistral-large-2512
Max outputMaximum output tokens per request.
210K
Knowledge cutoffLast date of training data.
—
ToolsTool calls on OpenRouter.
Yes
JSON outputOutput follows a given JSON schema.
Yes
Cache read / 1MUS dollars per 1M cached tokens.
$0.05
WeightsLink to the published weights.
Hugging Face (opens in a new tab) ↗
RetiresDate OpenRouter removes the model.
—

Where to run

APIs that serve this model.
2 providers
ProviderTypeMaker, cloud, host or router.InputOutputContext
Amazon BedrockCloud$0.50$1.50262K
OpenRouterRouter$0.50$1.50262K

US dollars per 1M tokens, list price. From models.dev (opens in a new tab) (MIT).

Run locally

675B parameters
16-bit
1418.5 GB
8-bit
754.0 GB
4-bit
430.7 GB
Weights
Hugging Face (opens in a new tab) ↗

Memory need with a short context. The 4-bit copy needs more than 128 GB. Compare local models

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.4 counted29.7202 / 218
CodingCode writing and repair.4 counted38.2184 / 209
ReasoningLogic problems and puzzles.2 counted34.4193 / 227
MultimodalTasks with images and text.1 counted34.5—
KnowledgeFacts and expert knowledge.4 counted38.9201 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted22.9—
MathMath problems.1 counted34.3—

Results

17 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
Artificial Analysis Omniscience AccuracyKnowledge25.0%105 / 164
MMLU ProKnowledge79.8% ±0.592 / 135
Artificial Analysis GPQA DiamondKnowledge68.0%137 / 199
AIMEMath42.9% ±0.966 / 95
Critical Physics TasksReasoning0.0%150 / 214
GPQA Diamond · Vals AIKnowledge68.4% ±2.498 / 135
GDPval-AA normalizedAgentic0.0%101 / 133
LiveCodeBenchCoding55.3% ±1.2112 / 142
Artificial Analysis Long Context ReasoningReasoning36.0%174 / 217
IOI v1Coding4.0% ±1.549 / 61
MMMU ProMultimodal66.2% ±1.174 / 92
Terminal-Bench 1.0Agentic21.3% ±4.638 / 47
τ²-Bench Tool-Agent-User EvaluationAgentic24.6%122 / 149
Artificial Analysis IFBenchInstruction36.2%120 / 146
Artificial Analysis MMMU-ProMultimodal55.7%95 / 113
Artificial Analysis SciCodeCoding36.6%98 / 114
Artificial Analysis Humanity's Last ExamKnowledge4.2%194 / 222
SWE-benchCoding41.4% ±2.284 / 87
Terminal-Bench 2.0 · Vals AIAgentic9.0% ±3.064 / 66
TaxEval v2Finance73.1% ±0.952 / 143
CaseLaw v2Legal61.4% ±1.021 / 53
CorpFin v2Finance61.0% ±1.060 / 131
LegalBenchLegal79.1% ±0.598 / 149
MedQAHealthcare82.2% ±0.367 / 95
MortgageTaxFinance52.1% ±1.082 / 97
Finance Agent v1.1Finance18.0% ±2.246 / 51
MGSMMultilingual85.4% ±1.867 / 74

Sources

BenchLMCC BY-NC 4.0OpenRouterTerms restrict reuseVals AINo licence statedmodels.devMIT

History

5 releases
Index35.1 → 34.3
23 Oct 202612 Nov 2026
Output / 1M$1.50 → $1.50
23 Oct 202612 Nov 2026

Questions

How much does Mistral Large 3 cost?
$0.50 per 1M input tokens, $1.50 per 1M output tokens.
How good is Mistral Large 3?
Its index is 34.3 out of 100. That is place 238 of 283 ranked models.
What is Mistral Large 3 best at?
Knowledge: 38.9 out of 100, place 201 of 268.
How fast is Mistral Large 3?
61 output tokens per second, 0.7 s to the first token.
What is the context window of Mistral Large 3?
262K tokens.
Is Mistral Large 3 open source?
Yes. Its weights are published.
When was Mistral Large 3 released?
December 2025.

Badge

Mistral Large 3 on Inference360: place 238 of 283, index 34.3
HTML
<a href="https://inference360.ai/models/mistral-large-3"><img src="https://inference360.ai/badges/mistral-large-3.svg" alt="Mistral Large 3 on Inference360: place 238 of 283, index 34.3" height="20"></a>
Markdown
[![Mistral Large 3 on Inference360: place 238 of 283, index 34.3](https://inference360.ai/badges/mistral-large-3.svg)](https://inference360.ai/models/mistral-large-3)

Compare with

Llama 4 Maverick34.4GPT-4.1 mini34.5Magistral Medium33.7Celeris-133.7Magistral Small31.4

Lower price

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

More from Mistral

All models
Ministral 3 14BUnrankedMinistral 3 14B (Reasoning)UnrankedMinistral 3 3BUnrankedMinistral 3 3B (Reasoning)UnrankedMinistral 3 8BUnranked