Mistral AI model product
Mistral Large 4
Mistral Large 4 is a 1T-parameter multimodal MoE model from Mistral AI for reasoning, coding, and agentic work.
Updated Oct 23, 2026. Default version: Mistral Large 4
LLMBoard Score74.4Mistral Large 4
Coverage20%12 benchmark families
Context window1MTokens
Official input price$1.36Mistral API
Mistral Large 4 Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
Mistral Large 4 Benchmark Results
Benchmark scores for Mistral Large 4.
Sort by
Benchmark | Score | Rank | Participants | Percentile | Evidence | Evaluated |
|---|
| BenchmarkSciCode | Score91.80% | Rank01 | Participants31 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkCoding Agent Index | Score49.80% | Rank01 | Participants1 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkDense200 | Score42.00% | Rank01 | Participants1 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkFinch (FinWorkBench) | Score67.40% | Rank01 | Participants1 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkB3 AI Security Benchmark | Score93.30% | Rank01 | Participants1 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkKORABench | Score1.69 points | Rank01 | Participants1 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkSurge AI Coding Quality | Score3.74 points | Rank01 | Participants1 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkAutomationBench | Score59.90% | Rank02 | Participants27 | Percentile96.15% | EvidenceC | Evaluated |
| BenchmarkCyBench | Score93.00% | Rank02 | Participants3 | Percentile50.00% | EvidenceC | Evaluated |
| BenchmarkSWE Atlas - Codebase QnA | Score59.40% | Rank02 | Participants7 | Percentile83.33% | EvidenceC | Evaluated |
| BenchmarkAA-Briefcase | Score1,393.00 points | Rank03 | Participants4 | Percentile33.33% | EvidenceC | Evaluated |
| BenchmarkHarvey's Legal Agent Benchmark | Score15.80% | Rank03 | Participants4 | Percentile33.33% | EvidenceC | Evaluated |
| BenchmarkChartQAPro | Score63.10% | Rank04 | Participants4 | Percentile0.00% | EvidenceC | Evaluated |
| BenchmarkFinance Agent v2 | Score54.70% | Rank06 | Participants29 | Percentile82.14% | EvidenceC | Evaluated |
| BenchmarkLiveBench math (2026-06-25) | Score93.58 score | Rank10 | Participants43 | Percentile78.57% | EvidenceB | EvaluatedN/A |
| BenchmarkCyberGym | Score82.00% | Rank12 | Participants23 | Percentile50.00% | EvidenceC | Evaluated |
| BenchmarkGDP.pdf | Score18.60% | Rank14 | Participants14 | Percentile0.00% | EvidenceC | Evaluated |
| BenchmarkLM Arena Agent Bash Recovery Steps | Score1.78% | Rank20 | Participants44 | Percentile55.81% | EvidenceA | Evaluated |
| BenchmarkLM Arena Agent Steerability | Score0.35% | Rank20 | Participants44 | Percentile55.81% | EvidenceA | Evaluated |
| BenchmarkLM Arena Agent Task Outcome Explicit | Score0.91% | Rank20 | Participants44 | Percentile55.81% | EvidenceA | Evaluated |
| BenchmarkTerminal-Bench 4.0 | Score28.30% | Rank21 | Participants29 | Percentile28.57% | EvidenceC | Evaluated |
| BenchmarkLiveBench reasoning (2026-06-25) | Score83.94 score | Rank24 | Participants43 | Percentile45.24% | EvidenceB | EvaluatedN/A |
| BenchmarkLM Arena Agent Bash Reliability | Score1.06% | Rank24 | Participants44 | Percentile46.51% | EvidenceA | Evaluated |
| BenchmarkLM Arena Agent Bash Reliability No Prereq | Score2.14% | Rank25 | Participants45 | Percentile45.45% | EvidenceA | Evaluated |
| BenchmarkLM Arena Agent Praise Complaint | Score-4.40% | Rank28 | Participants44 | Percentile37.21% | EvidenceA | Evaluated |
| BenchmarkLM Arena Agent Leaderboard | Score-4.55% | Rank31 | Participants44 | Percentile30.23% | EvidenceA | Evaluated |
| BenchmarkLM Arena Webdev | Score1,542.52 rating | Rank31 | Participants107 | Percentile71.70% | EvidenceA | Evaluated |
| BenchmarkLiveBench instruction (2026-06-25) | Score64.82 score | Rank31 | Participants43 | Percentile28.57% | EvidenceB | EvaluatedN/A |
| BenchmarkDeepSWE 1.1 | Score61.70% | Rank32 | Participants51 | Percentile38.00% | EvidenceC | Evaluated |
| BenchmarkLM Arena Vision | Score1,272.17 rating | Rank33 | Participants113 | Percentile71.43% | EvidenceA | Evaluated |
Mistral Large 4 Arena Results
Preference and agent-evaluation results for the default version.
Sort by
Arena | Category | Rank | Rating / score | Votes | Observations | Result date |
|---|
| Arenatext | Categoryfrench | Rank18 | Rating / score1,492.29 | Votes395 | ObservationsN/A | Result date |
| Arenaagent bash recovery steps | Categoryoverall | Rank20 | Rating / score0.02 | VotesN/A | Observations6K | Result date |
| Arenaagent steerability | Categoryoverall | Rank20 | Rating / score0.00 | VotesN/A | Observations40.5K | Result date |
| Arenaagent task outcome explicit | Categoryoverall | Rank20 | Rating / score0.01 | VotesN/A | Observations4.9K | Result date |
| Arenatext style control | Categoryfrench | Rank24 | Rating / score1,494.48 | Votes395 | ObservationsN/A | Result date |
| Arenaagent bash reliability | Categoryoverall | Rank24 | Rating / score0.01 | VotesN/A | Observations89.7K | Result date |
| Arenaagent bash reliability no prereq | Categoryoverall | Rank25 | Rating / score0.02 | VotesN/A | Observations43K | Result date |
| Arenavision | Categoryenglish | Rank27 | Rating / score1,283.20 | Votes355 | ObservationsN/A | Result date |
| Arenaagent praise complaint | Categoryoverall | Rank28 | Rating / score-0.04 | VotesN/A | Observations2.1K | Result date |
| Arenawebdev | Categorywebdev-react | Rank28 | Rating / score1,558.56 | Votes1,783 | ObservationsN/A | Result date |
| Arenaagent | Categoryoverall | Rank31 | Rating / score-0.05 | VotesN/A | Observations1M | Result date |
| Arenavision | Categoryocr | Rank31 | Rating / score1,284.36 | Votes671 | ObservationsN/A | Result date |
| Arenawebdev | Categoryoverall | Rank31 | Rating / score1,542.52 | Votes2,832 | ObservationsN/A | Result date |
| Arenawebdev | Categorywebdev | Rank31 | Rating / score1,542.52 | Votes2,832 | ObservationsN/A | Result date |
| Arenavision | Categorydiagram | Rank33 | Rating / score1,287.84 | Votes243 | ObservationsN/A | Result date |
| Arenavision | Categoryoverall | Rank33 | Rating / score1,272.17 | Votes1,008 | ObservationsN/A | Result date |
| Arenavision style control | Categoryenglish | Rank35 | Rating / score1,262.87 | Votes355 | ObservationsN/A | Result date |
| Arenavision style control | Categoryocr | Rank37 | Rating / score1,270.16 | Votes671 | ObservationsN/A | Result date |
| Arenatext | Categorymath | Rank38 | Rating / score1,468.79 | Votes488 | ObservationsN/A | Result date |
| Arenavision style control | Categorydiagram | Rank38 | Rating / score1,281.57 | Votes243 | ObservationsN/A | Result date |
| Arenavision style control | Categoryoverall | Rank39 | Rating / score1,251.94 | Votes1,008 | ObservationsN/A | Result date |
| Arenatext | Categoryspanish | Rank40 | Rating / score1,455.14 | Votes320 | ObservationsN/A | Result date |
| Arenaagent tool hallucination | Categoryoverall | Rank44 | Rating / score-0.18 | VotesN/A | Observations876K | Result date |
| Arenatext | Categorycoding | Rank44 | Rating / score1,478.33 | Votes2,115 | ObservationsN/A | Result date |
| Arenatext | Categoryindustry business and management and financial operations | Rank44 | Rating / score1,439.32 | Votes1,568 | ObservationsN/A | Result date |
| Arenatext | Categoryjapanese | Rank44 | Rating / score1,415.17 | Votes194 | ObservationsN/A | Result date |
| Arenatext style control | Categorymath | Rank45 | Rating / score1,468.30 | Votes488 | ObservationsN/A | Result date |
| Arenawebdev | Categorywebdev-html | Rank45 | Rating / score1,487.05 | Votes374 | ObservationsN/A | Result date |
| Arenatext factuality | Categorycoding | Rank49 | Rating / score1,503.28 | Votes2,096 | ObservationsN/A | Result date |
| Arenatext | Categorychinese | Rank50 | Rating / score1,491.97 | Votes825 | ObservationsN/A | Result date |
Mistral Large 4 Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- $1.36 input, $4.18 output per 1M
- Official provider
- Mistral
- Lowest third-party
- From $0.68 input, $2.09 output per 1M via Requesty
- Tracked offerings
- 13
Sort by
Provider | Provider model ID | Region | Input / 1M | Output / 1M | Context | Updated |
|---|
| ProviderRequesty | Provider model IDmistral-large-4 | Regionglobal | Input / 1M$0.68 | Output / 1M$2.09 | Context524.3K | Updated |
| ProviderNanoGPT | Provider model IDmistralai/mistral-large-4 | Regionglobal | Input / 1M$0.68 | Output / 1M$2.09 | Context524.3K | Updated |
| ProviderOllama Cloud | Provider model IDmistral-large-4 | Regionglobal | Input / 1M$0.68 | Output / 1M$2.09 | Context1M | Updated |
| ProviderOpenCode | Provider model IDmistral-large-4 | Regionglobal | Input / 1M$0.68 | Output / 1M$2.09 | Context524.3K | Updated |
| ProviderStandard Compute | Provider model IDmistralai/mistral-large-4-0 | Regionglobal | Input / 1M$0.68 | Output / 1M$2.09 | Context1M | Updated |
| ProviderCortecs | Provider model IDmistral-large-4-0 | Regionglobal | Input / 1M$0.765 | Output / 1M$2.35 | Context1M | Updated |
| ProviderOpenRouter | Provider model IDmistralai/mistral-large-4-0 | Regionglobal | Input / 1M$1.36 | Output / 1M$4.18 | Context1M | Updated |
| ProviderVercel AI Gateway | Provider model IDmistral/mistral-large-4 | Regionglobal | Input / 1M$1.36 | Output / 1M$4.18 | Context524.3K | Updated |
| ProviderMistral | Provider model IDmistral-large-4 | Regionglobal | Input / 1M$1.36 | Output / 1M$4.18 | Context1M | Updated |
| ProviderLLM Gateway | Provider model IDmistral/mistral-large-4 | Regionglobal | Input / 1M$1.36 | Output / 1M$4.18 | Context1M | Updated |
| ProviderDevPass (LLM Gateway) | Provider model IDmistral-large-4 | Regionglobal | Input / 1M$1.36 | Output / 1M$4.18 | Context1M | Updated |
| ProviderCommandCode | Provider model IDmistral/mistral-large-4 | Regionglobal | Input / 1M$1.36 | Output / 1M$4.18 | Context524.3K | Updated |
| ProviderLowRouter | Provider model IDauto/mistralai/mistral-large-4 | Regionglobal | Input / 1M$1.44 | Output / 1M$4.42 | Context524.3K | Updated |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
Mistral Large 4 Runtime Performance
Provider-specific output speed and catalog latency for Mistral Large 4. Runtime does not affect the capability score.
No runtime data
No provider-specific speed or latency record is linked to the default version yet.
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
Mistral Large 4 Specifications
Technical details for the model's default version.
- Version
- Mistral Large 4
- Released
- Oct 6, 2026
- Knowledge cutoff
- Unknown
- Parameters
- 1.1T
- Context window
- 1M
- Max output
- N/A
- Inputs
- image, text
- Outputs
- text
- Open weights
- No
- License
- Proprietary
Mistral Large 4 Versions
Available versions of this model. The score column identifies the version used in the overall ranking.
Sort by
Version | Released | LLMBoard | Parameters | Context | Max output | Open weights | License |
|---|
| VersionMistral Large 4 | Released | LLMBoard74.44 | Parameters1.1T | Context1M | Max outputN/A | Open weightsNo | LicenseProprietary |
Models similar to Mistral Large 4
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
MA
Mistral Medium 3.5
Mistral AI
MA
Mistral Small 4
Mistral AI
MA
Ministral 3
Mistral AI
MA
Devstral Medium
Mistral AI
MA
Pixtral Large
Mistral AI
MA
Min istral 3
Mistral AI
What is Mistral Large 4?
Key information about Mistral Large 4 and its available data.
Mistral Large 4 is a 1T-parameter multimodal mixture-of-experts (MoE) large language model from Mistral AI. It is intended for reasoning, coding, and agentic work.
Data as of 2026-10-22.
FAQ
Common questions about Mistral Large 4.
When was Mistral Large 4 released?
Mistral Large 4's default version was released on Oct 6, 2026.
How much does Mistral Large 4 cost?
Mistral Large 4's official API price is $1.36 per million input tokens and $4.18 per million output tokens via Mistral. The lowest tracked third-party offer starts at $0.68 input and $2.09 output via Requesty.
Who created Mistral Large 4?
Mistral Large 4 was created by Mistral AI.
What is the context window for Mistral Large 4?
The default version has a 1M token context window.
Is Mistral Large 4 open weight?
No. The default version is not marked as having publicly available weights.