geekradar
Int
Explore GeekRadar

More to explore

Good gear All topics About us Say hello
GeekRadar / The comparison desk
COMPARE.
AI models
mistral

Mistral Medium 3

AI model Released

API mode
Chat*
Context window
131,072 tokens
Maximum output
104,857 tokens
Open weights
No

Capability & preference

General capability · ECI

A cross-benchmark capability estimate. Index points, not a percentage; intervals show uncertainty.

Mistral Medium 3 134.0* Mistral Medium 3 · 95% interval 130.4–135.4

Human preference · Arena

Overall text preferences from head-to-head votes. Not factual accuracy or a task-success rate.

Mistral Medium 3 1370.1 mistral-medium-2505 · 95% interval 1365.5–1374.8 · 32,782 votes · 2026-10-15

Model performance

Radar task score: four tests, equal weight. An experimental summary of these tasks, not a universal intelligence rating.

Mistral Medium 3

Radar task score

Unrated

1/4 core tests · not enough evidence

1 benchmark results for this configuration

Scientific reasoning

GPQA Diamond · 25%

Expert-written science questions; strict multiple-choice scoring. Chance accuracy is 25%.

Mistral Medium 3 59.5*% Approx. 95% interval: 54.0–65.1% · Older result

Novel problem solving

Mystery Game Puzzles · 25%

Best-next-move puzzles in an undisclosed game variant. A narrow reasoning test, not general intelligence.

Mistral Medium 3 Not available

Advanced mathematics

FrontierMath Tiers 1-3 · 25%

Private v2 mathematics problems with Python tools. Earlier versions and Tier 4 are excluded.

Mistral Medium 3 Not available

Factual answers

SimpleQA Verified · 25%

Correct factual answers under a guessing prompt, graded by a model. Not a hallucination or abstention score.

Mistral Medium 3 Not available

Bars show reported task scores. Whiskers are approximate 95% intervals where standard errors are available. Different reasoning budgets are not equal-cost tests. Missing results are never zero.

Recorded results

Only the same benchmark, version, workload, scenario and unit share a row. Missing results are not zero. System results include the full submitted hardware and software configuration.

Arena text preference / chinesehuman preference · 2026-10-15 · rating

Mistral Medium 3

1,373.61 rating

Confidence interval: 1,360.66 – 1,386.55

2048 votes

Category
chinese
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / codinghuman preference · 2026-10-15 · rating

Mistral Medium 3

1,385.42 rating

Confidence interval: 1,377.43 – 1,393.41

5843 votes

Category
coding
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / creative_writinghuman preference · 2026-10-15 · rating

Mistral Medium 3

1,343.52 rating

Confidence interval: 1,334.03 – 1,353.00

4063 votes

Category
creative_writing
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / englishhuman preference · 2026-10-15 · rating

Mistral Medium 3

1,384.43 rating

Confidence interval: 1,378.53 – 1,390.33

15111 votes

Category
english
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / exclude_tieshuman preference · 2026-10-15 · rating

Mistral Medium 3

1,340.37 rating

Confidence interval: 1,334.00 – 1,346.75

22754 votes

Category
exclude_ties
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / experthuman preference · 2026-10-15 · rating

Mistral Medium 3

1,341.84 rating

Confidence interval: 1,328.41 – 1,355.26

1806 votes

Category
expert
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / frenchhuman preference · 2026-10-15 · rating

Mistral Medium 3

1,367.92 rating

Confidence interval: 1,341.62 – 1,394.22

496 votes

Category
french
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / germanhuman preference · 2026-10-15 · rating

Mistral Medium 3

1,377.12 rating

Confidence interval: 1,358.50 – 1,395.75

997 votes

Category
german
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / hard_promptshuman preference · 2026-10-15 · rating

Mistral Medium 3

1,364.50 rating

Confidence interval: 1,358.55 – 1,370.45

13650 votes

Category
hard_prompts
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / hard_prompts_englishhuman preference · 2026-10-15 · rating

Mistral Medium 3

1,373.71 rating

Confidence interval: 1,365.94 – 1,381.47

6365 votes

Category
hard_prompts_english
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-15 · rating

Mistral Medium 3

1,351.80 rating

Confidence interval: 1,343.12 – 1,360.47

5094 votes

Category
industry_business_and_management_and_financial_operations
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / industry_entertainment_and_sports_and_mediahuman preference · 2026-10-15 · rating

Mistral Medium 3

1,335.08 rating

Confidence interval: 1,326.96 – 1,343.20

5602 votes

Category
industry_entertainment_and_sports_and_media
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / industry_legal_and_governmenthuman preference · 2026-10-15 · rating

Mistral Medium 3

1,383.77 rating

Confidence interval: 1,370.97 – 1,396.56

2118 votes

Category
industry_legal_and_government
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / industry_life_and_physical_and_social_sciencehuman preference · 2026-10-15 · rating

Mistral Medium 3

1,381.00 rating

Confidence interval: 1,372.67 – 1,389.33

5635 votes

Category
industry_life_and_physical_and_social_science
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / industry_mathematicalhuman preference · 2026-10-15 · rating

Mistral Medium 3

1,354.10 rating

Confidence interval: 1,341.92 – 1,366.29

2192 votes

Category
industry_mathematical
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / industry_medicine_and_healthcarehuman preference · 2026-10-15 · rating

Mistral Medium 3

1,381.12 rating

Confidence interval: 1,367.51 – 1,394.72

1971 votes

Category
industry_medicine_and_healthcare
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / industry_software_and_it_serviceshuman preference · 2026-10-15 · rating

Mistral Medium 3

1,385.17 rating

Confidence interval: 1,378.64 – 1,391.70

10552 votes

Category
industry_software_and_it_services
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / industry_writing_and_literature_and_languagehuman preference · 2026-10-15 · rating

Mistral Medium 3

1,351.87 rating

Confidence interval: 1,344.36 – 1,359.38

7177 votes

Category
industry_writing_and_literature_and_language
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / instruction_followinghuman preference · 2026-10-15 · rating

Mistral Medium 3

1,338.50 rating

Confidence interval: 1,331.36 – 1,345.63

7948 votes

Category
instruction_following
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / japanesehuman preference · 2026-10-15 · rating

Mistral Medium 3

1,315.88 rating

Confidence interval: 1,295.88 – 1,335.87

916 votes

Category
japanese
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / koreanhuman preference · 2026-10-15 · rating

Mistral Medium 3

1,304.39 rating

Confidence interval: 1,280.62 – 1,328.17

639 votes

Category
korean
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / longer_queryhuman preference · 2026-10-15 · rating

Mistral Medium 3

1,360.14 rating

Confidence interval: 1,352.16 – 1,368.12

5928 votes

Category
longer_query
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / mathhuman preference · 2026-10-15 · rating

Mistral Medium 3

1,352.76 rating

Confidence interval: 1,340.80 – 1,364.73

2232 votes

Category
math
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / multi_turnhuman preference · 2026-10-15 · rating

Mistral Medium 3

1,383.54 rating

Confidence interval: 1,375.32 – 1,391.75

5716 votes

Category
multi_turn
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / non_englishhuman preference · 2026-10-15 · rating

Mistral Medium 3

1,353.15 rating

Confidence interval: 1,347.59 – 1,358.72

17663 votes

Category
non_english
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / overallhuman preference · 2026-10-15 · rating

Mistral Medium 3

1,370.15 rating

Confidence interval: 1,365.46 – 1,374.83

32782 votes

Category
overall
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / polishhuman preference · 2026-10-15 · rating

Mistral Medium 3

1,370.57 rating

Confidence interval: 1,358.92 – 1,382.22

2886 votes

Category
polish
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / russianhuman preference · 2026-10-15 · rating

Mistral Medium 3

1,359.42 rating

Confidence interval: 1,348.54 – 1,370.30

2761 votes

Category
russian
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference
Arena text preference / spanishhuman preference · 2026-10-15 · rating

Mistral Medium 3

1,363.97 rating

Confidence interval: 1,338.92 – 1,389.03

567 votes

Category
spanish
Evaluated configuration
mistral-medium-2505
Evaluation type
human preference

Run it locally

Inference · memory estimate

Mistral Medium 3

Closed weights. No supported local deployment or GPU requirement is published.

API cost

USD / month

Standard text-token estimate. Excludes taxes, caching, batch discounts, reasoning and multimodal surcharges. Provider limits and regional availability may differ.

Mistral Medium 3

Mistral Medium 3 / Pricing tiers 1
Provider / scopeRateCurrencyAmount / unit
Google Vertex AIMistral Medium 3Standard InputUSD0.4 / 1M tokens
OutputUSD2 / 1M tokens

Capabilities

Context window
131,072 tokens
Maximum input
262,144 tokens*
Maximum output
104,857 tokens
Reasoning
No
Tool calling
Yes
Structured output
Yes*
Open weights
No
Knowledge cutoff
2025-03-31
API mode
Chat*
Image input
Yes*
Tool choice
Yes*
Model license
Proprietary

Input modalities

  • text
  • image
  • pdf

Output modalities

  • text

System configuration

Detailed specifications

Specifications · 15 fields
cache read input token cost
1.5e-7*
Scope and conditions
native_context
provider
mistral*
mode
chat*
full_native_model_id
mistral/mistral-medium-3*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
input cost per token
1.5e-6*
Scope and conditions
native_context
provider
mistral*
mode
chat*
full_native_model_id
mistral/mistral-medium-3*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
litellm provider
mistral*
Scope and conditions
native_context
provider
mistral*
mode
chat*
full_native_model_id
mistral/mistral-medium-3*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
max input tokens
262,144*
Scope and conditions
native_context
provider
mistral*
mode
chat*
full_native_model_id
mistral/mistral-medium-3*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
max output tokens
262,144*
Scope and conditions
native_context
provider
mistral*
mode
chat*
full_native_model_id
mistral/mistral-medium-3*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
max tokens
262,144*
Scope and conditions
native_context
provider
mistral*
mode
chat*
full_native_model_id
mistral/mistral-medium-3*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
mode
chat*
Scope and conditions
native_context
provider
mistral*
mode
chat*
full_native_model_id
mistral/mistral-medium-3*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
output cost per token
7.5e-6*
Scope and conditions
native_context
provider
mistral*
mode
chat*
full_native_model_id
mistral/mistral-medium-3*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
reasoning effort levels
  • none*
  • high*
Scope and conditions
native_context
provider
mistral*
mode
chat*
full_native_model_id
mistral/mistral-medium-3*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supports assistant prefill
Yes*
Scope and conditions
native_context
provider
mistral*
mode
chat*
full_native_model_id
mistral/mistral-medium-3*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supports function calling
Yes*
Scope and conditions
native_context
provider
mistral*
mode
chat*
full_native_model_id
mistral/mistral-medium-3*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supports reasoning
Yes*
Scope and conditions
native_context
provider
mistral*
mode
chat*
full_native_model_id
mistral/mistral-medium-3*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supports response schema
Yes*
Scope and conditions
native_context
provider
mistral*
mode
chat*
full_native_model_id
mistral/mistral-medium-3*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supports tool choice
Yes*
Scope and conditions
native_context
provider
mistral*
mode
chat*
full_native_model_id
mistral/mistral-medium-3*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*
supports vision
Yes*
Scope and conditions
native_context
provider
mistral*
mode
chat*
full_native_model_id
mistral/mistral-medium-3*
pricing_scope
Literal source field names and values retained; provider, region, cache/batch/threshold conditions and per-unit meaning must be checked. max_tokens is not automatically a context limit.*

* Provisional data