Skip to main content
BenchLM
Data

Data as of November 21, 2026 · How the score is built

Mistral Large 4

CurrentOpen WeightHybrid

Decision readingMistral Large 4 scores 56.9 out of 100 and ranks #68 of 215. This profile shows 19 source-displayable benchmark rows; its strongest eligible category is Coding at #43. API pricing is $0.7 input and $2.1 output per million tokens, with cached input at $0.07.

Email me when its price, rank or availability changes.

Capability

56.9/100

field median 48.1#68 of 215 ranked models

Public

#68of 215

Verified #60 of 145

Price

$0.70input / $2.10 output

input median $0.60cached $0.070 · blended $1.40

Speed

57tok/s

field median 93.5 tok/sFirst token 84.04 s

Context

1Mtokens

field median 256,000Maximum output length is tracked separately

Decision snapshot

Where this model sits in the ranked field, its strongest published evidence, and what to validate before choosing it. The marker compares it with the current ranked catalog, not with an absolute quality threshold.

Strongest published evidence

Coding ranks #43. Particularly well-suited for software development and code generation tasks.

Validate before choosing

19 published rows leave some tracked benchmark slots empty. Knowledge is its lowest eligible category at #96.

Source-linked · 19 displayable benchmark rows

Mistral Large 4 API pricing

Mistral Large 4 costs $0.70 per million input tokens and $2.10 per million output tokens on the Mistral API. Cached input is $0.07 per million.

Mistral Large 4 API pricing, per million tokens
Token typePrice per 1M tokens
Input$0.70
Cached input$0.07
Output$2.10

At these rates, a request with 10,000 input tokens and 1,000 output tokens costs $0.0091 without caching.

On blended price (the average of the input and output rates), 9 of the 122 other ranked models with a published API price cost the same or less and score higher. See where it sits on the chart

Pricing notes: Mistral’s October 29, 2026 open-weight announcement prices Mistral Large 4 at $0.70 input, $0.07 cached input, and $2.10 output per million tokens, starting that day. At 14:20 UTC the same day, Mistral’s model page, API pricing page, and OpenRouter still listed the earlier $1.36/$0.14/$4.18 list price (the October 6 preview sale price was $0.68/$0.07/$2.09).

Rates from Mistral AI: Mistral Large 4 is now open weight. Price registry updated November 20, 2026.

Is Mistral Large 4 right for your job?

Its place on the LLM Selector’s job rankings, the same order the job pages publish. Describe your own job to see where it stands for that.

Describe the task and constraints. This selector recommends models; it cannot carry out the task or access your files and accounts.

We keep what you type here, drafts included, for 180 days, and after that only the job and requirements read from it. Nothing is tied to your IP address or account. Leave out names and other personal details. Privacy

Category score record

Scores and ranks appear only where published evidence can be displayed. The table keeps the score, weight, cohort, and evidence state together.

Category scores, ranks, weighting, benchmark coverage, and evidence status
CategoryScore
AgenticRank #74 of 179Percentile 59thWeight 22%8 benchmarksVerified
56.2
CodingRank #43 of 179Percentile 76thWeight 20%2 benchmarksVerified
63.2
ReasoningRank #57 of 161Percentile 65thWeight 17%1 benchmarkVerified
56.0
MultimodalRank #77 of 119Percentile 36thWeight 12%Not listedUnlisted sources
50.8
KnowledgeRank #96 of 158Percentile 39thWeight 12%Not listedUnlisted sources
48.1
MultilingualWeight 7%0 benchmarksNot measured
Not measured
Inst. FollowingRank #62 of 143Percentile 57thWeight 5%Not listedUnlisted sources
53.5
MathRank #48 of 125Percentile 62ndWeight 5%Not listedUnlisted sources
55.5

19 of 937 tracked benchmark slots have displayable evidence · bars run 0–100

Coverage details

How much of this is verified

Coverage is split by category so a strong number never hides a thin evidence base. Verified means the row is tied to a published source; provisional rows remain visible but separate.

  1. Agentic8/8 verified
  2. Coding2/2 verified
  3. Reasoning1/1 verified
  4. MultimodalNot measured
  5. KnowledgeNot measured
  6. MultilingualNot measured
  7. Inst. FollowingNot measured
  8. MathNot measured
Verified sourceProvisionalNot measured

Capability shape

Each axis shows percentile within that category’s eligible cohort. The comparison outline is the median of the six nearest public-score peers; a collapsed vertex means the category is not rank-eligible.

Mistral Large 4 category percentile values

  • Agentic59th percentile
  • Coding76th percentile
  • Reasoning65th percentile
  • Multimodal36th percentile
  • Knowledge39th percentile
  • MultilingualNot eligible
  • Instruction following57th percentile
  • Math62nd percentile

The dashed outline is median of 6 nearest peers.

Eligible category ranks

  1. Agentic#74/179
  2. Coding#43/179
  3. Reasoning#57/161
  4. Multimodal#77/119
  5. Knowledge#96/158
  6. MultilingualNot ranked
  7. Inst. Following#62/143
  8. Math#48/125
Top decileTop quartileMid-fieldNot eligible

Benchmark ledger

Every result says who ran the test. For this model, Mistral ran 5 of 19. Coding opens by default. The tick marks the best source-verified result in the catalog; provisional leaders do not set the reference.

Overall rank
#68
of 215 on the leaderboard
Reported by Mistral
5
Shown with that label
Run by a benchmark
14
The benchmark’s own run
Independent runs
0
None yet
Coding2 rows
Coding benchmark results, best verified comparison, who ran each test, and whether it counts toward a rank
Vibe Code BenchVibe Code Bench v1.178.40%

Best verified: Gemini 4 Argon · 91.90%

Gap13.5 behindRun by Vibe Code BenchShown only
FrontierCode 2.019.7%

Best verified: Claude Fable 5.2 · 50.2%

Gap30.5 behindRun by FrontierCode 2.0Shown only
Agentic8 rows
Agentic benchmark results, best verified comparison, who ran each test, and whether it counts toward a rank
Terminal-Bench 4.127.60%

Best verified: Claude Fable 5.2 · 65.80%

Gap38.2 behindRun by Terminal-Bench 4.1Counts4% ref. weight
BrowseComp84.5%

Best verified: Atria Dawn Preview · 92.5%

Gap8 behindReported by MistralShown only
FinchFinch (FinWorkBench): Finance & Accounting across Spreadsheet-Centric Enterprise Workflows73.3%

Best verified: Mistral Large 4 · 73.3%

GapHighest listedReported by MistralShown only
b3 Agent Securityb3: Backbone Breaker Benchmark97.1%

Best verified: Mistral Large 4 · 97.1%

GapHighest listedReported by MistralShown only
Cyber Misuse RefusalRefusal of harmful cyber requests (JailbreakBench, StrongREJECT, AgentHarm)96.6%

Best verified: Mistral Large 4 · 96.6%

GapHighest listedReported by MistralShown only
Cybench93.0%

Best verified: Mistral Large 4 · 93.0%

GapHighest listedReported by MistralShown only
Finance Agent v254.7%

Best verified: Gemini 4 Argon · 65.4%

Gap10.7 behindRun by Finance Agent v2Shown only
CWE-benchCWE-bench v245.6%

Best verified: Claude Fable 5.2 · 75.0%

Gap29.4 behindRun by CWE-benchShown only
Reasoning1 row
Reasoning benchmark results, best verified comparison, who ran each test, and whether it counts toward a rank
STEMBenchSTEMBench120: six-domain scientific research progress2.50

Best verified: GPT-6 Astra · 2.81

Gap0.3 behindRun by STEMBenchCounts4% ref. weight
External signals8 rows
External signals benchmark results, best verified comparison, who ran each test, and whether it counts toward a rank
FrontierPhysicsEnd-to-End Physics Research Benchmark for AI Agents0.0%

Best verified: Claude Fable 5.2 · 33.3%

Gap33.3 behindRun by FrontierPhysicsShown only
NYT ConnectionsNYT Connections LLM Benchmark (extended version)27.4%

Best verified: Gemini 4 Carbon · 99.3%

Gap71.9 behindRun by NYT ConnectionsShown only
StealthMarkStealthMark Intelligence Benchmarks (overall score)43.3

Best verified: GPT-6 Astra · 74.6

Gap31.3 behindRun by StealthMarkShown only
Chess PuzzlesChess Puzzles (Epoch AI)33.0%

Best verified: Gemini 4 Neon · 83.0%

Gap50 behindRun by Chess PuzzlesShown only
Mystery Game PuzzlesMystery Game Puzzles (Epoch AI)17.0%

Best verified: GPT-6.2 Astra · 95.0%

Gap78 behindRun by Mystery Game PuzzlesShown only
Furniture AssemblyFurniture Assembly Benchmark (Epoch AI)20.8%

Best verified: Claude Fable 5.2 · 90.0%

Gap69.2 behindRun by Furniture AssemblyShown only
MirrorCodeMirrorCode (Epoch AI and METR)40.0%

Best verified: Claude Sonnet 5.5 · 97.8%

Gap57.8 behindRun by MirrorCodeShown only
HoneyBenchHoneyBench v0.2 (Goodhart Labs)83.6%

Best verified: Claude Fable 5.2 · 11.9%

Gap71.7 behindRun by HoneyBenchShown only

Outlined bars are lab-reported. Filled bars are benchmark-run or independent.

What it costs to get this score

Published API price against the public score. The x-axis uses a log scale; the dashed path marks models that are not beaten by a cheaper, higher-scoring option. Price uses average of published input and output rates.

Current modelExplore all models

Mistral Large 4 · 56.9 score · $1.40 blended per million tokens

The chart opens on the current model. Scroll horizontally to inspect the full price axis.

30405060708090100$0.10$0.50$1$5$10$25↘ frontierMistral Large 4

Horizontal: blended price per million tokens, log scale · Vertical: public score

Published price history

This chart appears only when at least two dated first-party price records exist for the exact model. It shows rate changes, not an inferred cost trend.

0242026-102026-10Output priceInput price

2 dated first-party price records. The newest record lists $2.1 per million tokens.

Lineage

The sequence follows explicit supersedes links. Each score is estimated for that model; a relative can inform a sparse estimate but never sets a floor, so a newer release can score below an earlier one. Scores and prices remain blank when the corresponding public row or first-party rate is unavailable.

  1. Dec 2, 2025

    Mistral Large 3

    Score 37.7 · $0.5 / $1.5

  2. Oct 6, 2026 · you are here

    Mistral Large 4

    Score 56.9 · $0.7 / $2.1

Base entry

Spec sheet

Each documented value carries its source. Missing fields stay visible as not sourced or not published, rather than disappearing from the page.

Maximum output
Not sourced yet
Knowledge cutoff
Not sourced yet
Input modalities
text, imageMistral-Large-4-1T-A52B model card
Parameters
Not sourced yet
Availability
Open weights (mistralai/Mistral-Large-4-1T-A52B) under the Mistral Ecosystem Open-Weight license, a modified MIT license; businesses running a Model-as-a-Service offering need a separate commercial license. API ID mistral-large-4-0, with Low, Medium and High (default) reasoning modes.
Cloud regions
Not tracked yet
Lifecycle
Current
API capabilities
Tool calling, structured outputs, and batch support are not tracked yet
Prompt caching
Published at $0.070 per million cached input tokensMistral AI: Mistral Large 4 is now open weight
Self-host
Mistral Large 4 needs ~1128GB VRAM (8× NVIDIA H200 (141GB)).
Rate limits
Not tracked yet

How to read this profile

The visual layer above carries the decisions. These notes preserve the model, ranking, coverage, and family context behind the numbers.

Mistral Large 4 ranks #68 of 215 on the public leaderboard with a score of 56.9/100. Its source-verified position is #60 of 145.

Mistral Large 4 is a open weight model with a 1M context window. No explicit reasoning mode is documented in this profile.

Open weights (mistralai/Mistral-Large-4-1T-A52B) under the Mistral Ecosystem Open-Weight license, a modified MIT license; businesses running a Model-as-a-Service offering need a separate commercial license. API ID mistral-large-4-0, with Low, Medium and High (default) reasoning modes.

Its explicit predecessor is Mistral Large 3. 19 of 937 tracked benchmark slots currently have displayable evidence. Missing categories stay blank.

Its strongest eligible category is Coding at #43, while its lowest eligible position is Knowledge at #96. particularly well-suited for software development and code generation tasks.

Last updated November 21, 2026. Runtime fields remain blank until a sourced snapshot exists.

Deployment options

Self-host and provider-specific paths stay separate from benchmark evidence so operating constraints are visible before a score becomes the whole decision.

Where to run Mistral Large 4 by API

OpenRouter is the one host we track that publishes a per-token price for it. Prices read from each host’s own pages on November 16, 2026.

Inference hosts for Mistral Large 4 with input, output and cached-input prices per million tokens.
HostInput /1MOutput /1MCached /1MContext
OpenRouter$0.70$2.10$0.071.05M
Compare inference hosts

Self-host vs API cost

Estimates at 50,000 req/day · 1000 tokens/req average.

Mistral Large 4
API / mo$2,100
Self-host / mo$37,843
Break-evenNot reached at 70% load
Model the full break-even

Questions

How does Mistral Large 4 perform overall in AI benchmarks?

Mistral Large 4 ranks #68 out of 215 models on the public BenchAlign leaderboard, with a score of 56.9/100. Its evidence status is Supported, and this profile shows 19 source-displayable benchmark rows. The label describes evidence depth, not a provider quality claim; inspect category rows before choosing a workload.

Is Mistral Large 4 good for coding and programming?

Mistral Large 4 ranks #43 out of 179 eligible models for coding and programming, with a public category score of 63.2/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.

Is Mistral Large 4 good for reasoning and logic?

Mistral Large 4 ranks #57 out of 161 eligible models for reasoning and logic, with a public category score of 56/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.

Is Mistral Large 4 good for agentic tool use and computer tasks?

Mistral Large 4 ranks #74 out of 179 eligible models for agentic tool use and computer tasks, with a public category score of 56.2/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.

Is Mistral Large 4 open source?

Mistral Large 4 is an open-weight model from Mistral. Its weights can be downloaded for local or hosted deployment, subject to the published license. Open weight does not automatically mean open source: training data and training code may remain private, and commercial restrictions can still apply.

Does Mistral Large 4 have full benchmark coverage on BenchLM?

No. Mistral Large 4 currently has 36 source-displayable rows. The profile exposes published, non-generated evidence and leaves missing categories blank until an exact evaluation is available. Coverage describes how much was measured; it is not a penalty added to an individual benchmark result.

What is the context window size of Mistral Large 4?

Mistral Large 4 has a documented context window of 1M. That figure is the maximum combined prompt and retained-conversation space reported for this exact model; it is not the maximum output length. The profile keeps output limits separate because providers often publish those limits independently.

Compare Mistral Large 4 with every tracked model1033 comparisons