Model profile / Mistral AI

Mistral Medium 3.5

Mistral AI's open-weight multimodal model for agentic and coding use cases.

Release date unavailableOpen weightsProvider

Sources & details

Four areas of evidence

The capabilities behind the model.

How the indices work

Reasoning

102.4Index points

Mathematical and scientific problem solving.

This model│ Other models · 155 rated

Score and research range

Full-precision point: 102.42445862815174
Range: 99.81–105.04 · 90% Research Range
Scope: Conditional fuller R0 A

Score includes qualified prediction of missing evidence.

Coding

82.5Index points

Software engineering and programming.

This model│ Other models · 39 rated

Score and research range

Full-precision point: 82.52653494796996
Range: No qualified range available
Scope: Frozen Coding score under admitted comparison/reference/source reweighting

Score uses the admitted direct native evidence under the selected method.

General

No current rating

Factual reliability and instruction following.

Awaiting a qualified category score

Agentic

91.3Index points

Tool use, workflows and professional tasks.

This model│ Other models · 40 rated

Score and research range

Full-precision point: 91.32570308595463
Range: 89.73–92.71 · 90% Research Range
Scope: Agentic score under joint reference/source reweighting conditional on admitted W/X/P native outcomes

Score uses the admitted direct native evidence under the selected method.

10retained native records

10evidence families

Capabilities estimateQualified original evidence across four areas.

Current index snapshot · 2026-09-16. Each category uses its own scale; scores across categories are not directly comparable. Evidence counts describe the retained Capabilities inputs; a family can contain several tasks or configurations. A point lead does not establish superiority.

Performance in context

Capabilities, price & speed.

Explore every model

Mistral Medium 3.5 is highlighted wherever a compatible measurement is available.

Lumina Capabilities Index

Capabilities versus API price

The current Capabilities cohort, with this model highlighted.

Lumina

2025+ · 62 models · 1 highlighted

Chart loads as you explore

Lumina Capabilities Index

Capabilities versus output speed

The current Capabilities cohort, with this model highlighted.

Lumina

2025+ · 72 models · Developer API · 10K · 1 highlighted

Chart loads as you explore

Explore the detail

Sources & model details.

Original evidence, with its context.

Current index evidence 4Scores, research ranges and the sources behind each category.

Capabilities

Score uses the admitted direct native evidence under the selected method.

Exact checkpoint: mistral-medium-3-5
Method: methods-a4471cee84fd478d2ed3
Range interpretation: C1 native factor under conditional crossed source/family reweighting of admitted fixed-corpus reference and own portfolio
Sources: Artificial Analysis:Artificial Analysis · UNKNOWN:Cognition
Source cutoff: {"asOfMeaning":"Selected local candidate: retained historical measurements plus reviewed additions; execution dates remain source-native.","newIdentityRecovery":"2026-09-13T08:41:08.804270+00:00","publicProjectionDate":"2026-09-16T06:02:45.491398+00:00","scoring":"Inherited pinned baseline source vintages plus retained4A/4A.1 originals and SWE owner correction; heterogeneous execution dates, often unknown. Not a claim of simultaneous2026-09-13 execution."}

Explore this index

Coding

Score uses the admitted direct native evidence under the selected method.

Exact checkpoint: mistral-medium-3-5
Method: methods-a4471cee84fd478d2ed3
Range interpretation: Frozen Coding score under admitted comparison/reference/source reweighting
Sources: Artificial Analysis · Cognition
Source cutoff: {"asOfMeaning":"Selected local candidate: retained historical measurements plus reviewed additions; execution dates remain source-native.","newIdentityRecovery":"2026-09-13T08:41:08.804270+00:00","publicProjectionDate":"2026-09-16T06:02:45.491398+00:00","scoring":"Inherited pinned baseline source vintages plus retained4A/4A.1 originals and SWE owner correction; heterogeneous execution dates, often unknown. Not a claim of simultaneous2026-09-13 execution."}

Explore this index

Reasoning

Score includes qualified prediction of missing evidence.

Exact checkpoint: mistral-medium-3-5
Method: methods-a4471cee84fd478d2ed3
Range interpretation: Conditional fuller R0 A
Sources: aa
Source cutoff: {"asOfMeaning":"Selected local candidate: retained historical measurements plus reviewed additions; execution dates remain source-native.","newIdentityRecovery":"2026-09-13T08:41:08.804270+00:00","publicProjectionDate":"2026-09-16T06:02:45.491398+00:00","scoring":"Inherited pinned baseline source vintages plus retained4A/4A.1 originals and SWE owner correction; heterogeneous execution dates, often unknown. Not a claim of simultaneous2026-09-13 execution."}

Explore this index

Agentic

Score uses the admitted direct native evidence under the selected method.

Exact checkpoint: mistral-medium-3-5
Method: methods-a4471cee84fd478d2ed3
Range interpretation: Agentic score under joint reference/source reweighting conditional on admitted W/X/P native outcomes
Sources: Artificial Analysis
Source cutoff: {"asOfMeaning":"Selected local candidate: retained historical measurements plus reviewed additions; execution dates remain source-native.","newIdentityRecovery":"2026-09-13T08:41:08.804270+00:00","publicProjectionDate":"2026-09-16T06:02:45.491398+00:00","scoring":"Inherited pinned baseline source vintages plus retained4A/4A.1 originals and SWE owner correction; heterogeneous execution dates, often unknown. Not a claim of simultaneous2026-09-13 execution."}

Explore this index

Benchmark measurements 27Published values in each benchmark’s own units.

27 measurements

Mistral Medium 3.5 published benchmark measurements
Benchmark	Result	Published configuration	Source
Humanity's Last ExamSource-native version	12.79%	Artificial Analysis public model evaluation page; exact provider variant as listed on the page.Artificial Analysis evaluation harness	Artificial Analysis evaluations for mistral-medium-3-5 ↗Observed 2026-07-15 · checked 2026-07-15source-checked · reference-only
GPQA DiamondSource-native version	74.85%	Artificial Analysis public model evaluation page; exact provider variant as listed on the page.Artificial Analysis evaluation harness	Artificial Analysis evaluations for mistral-medium-3-5 ↗Observed 2026-07-15 · checked 2026-07-15source-checked · reference-only
MMMU-ProSource-native version	64.86%	Artificial Analysis public model evaluation page; exact provider variant as listed on the page.Artificial Analysis evaluation harness	Artificial Analysis evaluations for mistral-medium-3-5 ↗Observed 2026-07-15 · checked 2026-07-15source-checked · ranking-eligible
Terminal-Bench2.1	50.56%	Artificial Analysis public model evaluation page; exact provider variant as listed on the page.Artificial Analysis evaluation harness	Artificial Analysis evaluations for mistral-medium-3-5 ↗Observed 2026-07-15 · checked 2026-07-15source-checked · reference-only
GDPval-AA v2Source-native version	21.448%	Mistral Medium 3.5Artificial Analysis independent evaluation	GDPval-AA v2 Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15source-checked · ranking-eligible
AA Long Context ReasoningSource-native version	61%	Mistral Medium 3.5Artificial Analysis independent evaluation	AA Long Context Reasoning Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15source-checked · reference-only
AA-OmniscienceSource-native version	-36.317 index	Mistral Medium 3.5Artificial Analysis independent evaluation	AA-Omniscience Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15source-checked · reference-only
CritPtSource-native version	0%	Mistral Medium 3.5Artificial Analysis independent evaluation	CritPt Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15source-checked · ranking-eligible
τ²-Bench TelecomSource-native version	94.152%	Mistral Medium 3.5Artificial Analysis independent evaluation	τ²-Bench Telecom Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15source-checked · reference-only
τ³-BankingSource-native version	14.433%	Mistral Medium 3.5Artificial Analysis independent evaluation	τ³-Banking Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15source-checked · reference-only

Showing 10 of 27 measurements

Specifications & source recordsDocumented model details and original references.

Model ID / checkpoint: mistral-medium-3-5
Weight access: Open weights
Context window: 256,000
Maximum output: Not documented
Modalities: text, image
Details checked: 1 Aug 2026

Specialist task evidenceSeparate agent runs and execution conditions.

All models

Model profile / Mistral AI

Mistral Medium 3.5

Mistral AI's open-weight multimodal model for agentic and coding use cases.

Release date unavailableOpen weightsProvider

Sources & details

Four areas of evidence

The capabilities behind the model.

How the indices work

Reasoning

102.4Index points

Mathematical and scientific problem solving.

This model│ Other models · 155 rated

Score and research range

Full-precision point: 102.42445862815174
Range: 99.81–105.04 · 90% Research Range
Scope: Conditional fuller R0 A

Score includes qualified prediction of missing evidence.

Coding

82.5Index points

Software engineering and programming.

This model│ Other models · 39 rated

Score and research range

Full-precision point: 82.52653494796996
Range: No qualified range available
Scope: Frozen Coding score under admitted comparison/reference/source reweighting

Score uses the admitted direct native evidence under the selected method.

General

No current rating

Factual reliability and instruction following.

Awaiting a qualified category score

Agentic

91.3Index points

Tool use, workflows and professional tasks.

This model│ Other models · 40 rated

Score and research range

Full-precision point: 91.32570308595463
Range: 89.73–92.71 · 90% Research Range
Scope: Agentic score under joint reference/source reweighting conditional on admitted W/X/P native outcomes

Score uses the admitted direct native evidence under the selected method.

10retained native records

10evidence families

Capabilities estimateQualified original evidence across four areas.

Performance in context

Capabilities, price & speed.

Explore every model

Mistral Medium 3.5 is highlighted wherever a compatible measurement is available.

Lumina Capabilities Index

Capabilities versus API price

The current Capabilities cohort, with this model highlighted.

Lumina

2025+ · 62 models · 1 highlighted

Chart loads as you explore

Lumina Capabilities Index

Capabilities versus output speed

The current Capabilities cohort, with this model highlighted.

Lumina

2025+ · 72 models · Developer API · 10K · 1 highlighted

Chart loads as you explore

Explore the detail

Sources & model details.

Original evidence, with its context.

Current index evidence 4Scores, research ranges and the sources behind each category.

Capabilities

Score uses the admitted direct native evidence under the selected method.

Exact checkpoint: mistral-medium-3-5
Method: methods-a4471cee84fd478d2ed3
Range interpretation: C1 native factor under conditional crossed source/family reweighting of admitted fixed-corpus reference and own portfolio
Sources: Artificial Analysis:Artificial Analysis · UNKNOWN:Cognition
Source cutoff: {"asOfMeaning":"Selected local candidate: retained historical measurements plus reviewed additions; execution dates remain source-native.","newIdentityRecovery":"2026-09-13T08:41:08.804270+00:00","publicProjectionDate":"2026-09-16T06:02:45.491398+00:00","scoring":"Inherited pinned baseline source vintages plus retained4A/4A.1 originals and SWE owner correction; heterogeneous execution dates, often unknown. Not a claim of simultaneous2026-09-13 execution."}

Explore this index

Coding

Score uses the admitted direct native evidence under the selected method.

Exact checkpoint: mistral-medium-3-5
Method: methods-a4471cee84fd478d2ed3
Range interpretation: Frozen Coding score under admitted comparison/reference/source reweighting
Sources: Artificial Analysis · Cognition
Source cutoff: {"asOfMeaning":"Selected local candidate: retained historical measurements plus reviewed additions; execution dates remain source-native.","newIdentityRecovery":"2026-09-13T08:41:08.804270+00:00","publicProjectionDate":"2026-09-16T06:02:45.491398+00:00","scoring":"Inherited pinned baseline source vintages plus retained4A/4A.1 originals and SWE owner correction; heterogeneous execution dates, often unknown. Not a claim of simultaneous2026-09-13 execution."}

Explore this index

Reasoning

Score includes qualified prediction of missing evidence.

Exact checkpoint: mistral-medium-3-5
Method: methods-a4471cee84fd478d2ed3
Range interpretation: Conditional fuller R0 A
Sources: aa
Source cutoff: {"asOfMeaning":"Selected local candidate: retained historical measurements plus reviewed additions; execution dates remain source-native.","newIdentityRecovery":"2026-09-13T08:41:08.804270+00:00","publicProjectionDate":"2026-09-16T06:02:45.491398+00:00","scoring":"Inherited pinned baseline source vintages plus retained4A/4A.1 originals and SWE owner correction; heterogeneous execution dates, often unknown. Not a claim of simultaneous2026-09-13 execution."}

Explore this index

Agentic

Score uses the admitted direct native evidence under the selected method.

Exact checkpoint: mistral-medium-3-5
Method: methods-a4471cee84fd478d2ed3
Range interpretation: Agentic score under joint reference/source reweighting conditional on admitted W/X/P native outcomes
Sources: Artificial Analysis
Source cutoff: {"asOfMeaning":"Selected local candidate: retained historical measurements plus reviewed additions; execution dates remain source-native.","newIdentityRecovery":"2026-09-13T08:41:08.804270+00:00","publicProjectionDate":"2026-09-16T06:02:45.491398+00:00","scoring":"Inherited pinned baseline source vintages plus retained4A/4A.1 originals and SWE owner correction; heterogeneous execution dates, often unknown. Not a claim of simultaneous2026-09-13 execution."}

Explore this index

Benchmark measurements 27Published values in each benchmark’s own units.

27 measurements

Mistral Medium 3.5 published benchmark measurements
Benchmark	Result	Published configuration	Source
Humanity's Last ExamSource-native version	12.79%	Artificial Analysis public model evaluation page; exact provider variant as listed on the page.Artificial Analysis evaluation harness	Artificial Analysis evaluations for mistral-medium-3-5 ↗Observed 2026-07-15 · checked 2026-07-15source-checked · reference-only
GPQA DiamondSource-native version	74.85%	Artificial Analysis public model evaluation page; exact provider variant as listed on the page.Artificial Analysis evaluation harness	Artificial Analysis evaluations for mistral-medium-3-5 ↗Observed 2026-07-15 · checked 2026-07-15source-checked · reference-only
MMMU-ProSource-native version	64.86%	Artificial Analysis public model evaluation page; exact provider variant as listed on the page.Artificial Analysis evaluation harness	Artificial Analysis evaluations for mistral-medium-3-5 ↗Observed 2026-07-15 · checked 2026-07-15source-checked · ranking-eligible
Terminal-Bench2.1	50.56%	Artificial Analysis public model evaluation page; exact provider variant as listed on the page.Artificial Analysis evaluation harness	Artificial Analysis evaluations for mistral-medium-3-5 ↗Observed 2026-07-15 · checked 2026-07-15source-checked · reference-only
GDPval-AA v2Source-native version	21.448%	Mistral Medium 3.5Artificial Analysis independent evaluation	GDPval-AA v2 Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15source-checked · ranking-eligible
AA Long Context ReasoningSource-native version	61%	Mistral Medium 3.5Artificial Analysis independent evaluation	AA Long Context Reasoning Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15source-checked · reference-only
AA-OmniscienceSource-native version	-36.317 index	Mistral Medium 3.5Artificial Analysis independent evaluation	AA-Omniscience Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15source-checked · reference-only
CritPtSource-native version	0%	Mistral Medium 3.5Artificial Analysis independent evaluation	CritPt Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15source-checked · ranking-eligible
τ²-Bench TelecomSource-native version	94.152%	Mistral Medium 3.5Artificial Analysis independent evaluation	τ²-Bench Telecom Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15source-checked · reference-only
τ³-BankingSource-native version	14.433%	Mistral Medium 3.5Artificial Analysis independent evaluation	τ³-Banking Leaderboard ↗Observed 2026-07-15 · checked 2026-07-15source-checked · reference-only

Showing 10 of 27 measurements

Specifications & source recordsDocumented model details and original references.

Model ID / checkpoint: mistral-medium-3-5
Weight access: Open weights
Context window: 256,000
Maximum output: Not documented
Modalities: text, image
Details checked: 1 Aug 2026

Specialist task evidenceSeparate agent runs and execution conditions.

The capabilities behind the model.

Capabilities, price & speed.

Capabilities versus API price

Capabilities versus output speed

Sources & model details.

Capabilities

Coding

Reasoning

Agentic

More from Mistral AI.

Devstral 2

Mistral Small 4

Mistral Large 3

The capabilities behind the model.

Capabilities, price & speed.

Capabilities versus API price

Capabilities versus output speed

Sources & model details.

Capabilities

Coding

Reasoning

Agentic

More from Mistral AI.

Devstral 2

Mistral Small 4

Mistral Large 3