ModelBench
Mistral AI logo
Models/Mistral AI

Mistral Medium 3

Mistral model for multilingual chat, reasoning, and tool-assisted workflows

Source-linked mistral-medium Closed weights Released 2025-05-07
API data Report
Input modalitiesT
Output modalitiesT
Input price$0.4/1M tokens
Output price$2/1M tokens
Context window131.072K
Max output131.072K
Overview

Capabilities and specifications

× Reasoning No
× Tool calling No
× Structured output No
✓ Attachments Yes
✓ Vision input Yes
× Open weights No
Creator
Mistral AI
Model family
mistral-medium
Knowledge cutoff
2025-05
License
Not documented
Release date
2025-05-07
Model ID
mistral/mistral-medium-2505
Inference availability

Provider pricing and availability

Provider-specific identifiers, limits, and prices per million tokens.

Report a price
Providers offering Mistral Medium 3
ProviderProvider model IDContextInputOutputCache readCapabilitiesSource
Nvidia mistralai/mistral-medium-3-instruct 131.072K Not listed Not listed - Not listed Docs
EmpirioLabs AI mistral-medium-3 130K Not listed Not listed - ToolsJSON Docs
Mistral mistral-medium-2505 131.072K $0.4 $2 - Tools Docs
Merge Gateway mistral/mistral-medium-2505 128K $0.4 $2 $0.04 ToolsJSON Docs
Azure mistral-medium-2505 128K $0.4 $2 - Tools Docs
OpenRouter mistralai/mistral-medium-3 131.072K $0.4 $2 $0.04 ToolsJSON Docs
Azure Cognitive Services mistral-medium-2505 128K $0.4 $2 - Tools Docs
NanoGPT mistralai/mistral-medium-3 131.072K $0.4 $2 - Not listed Docs

Capability badges appear only when the provider catalog explicitly lists support.

Serving quality

Performance and reliability

Observed serving measurements stay separate from model specifications and are never estimated from price or model size.

Throughput Not reported

Generation speed; higher is faster.

Latency / TTFT Not reported

Time until the first output token arrives.

Uptime Not reported

Successful requests over a measured period.

Performance varies by inference provider, region, and load. Provider-level measurements appear only when a source reports them.

Published evaluations

Benchmarks

Raw results remain attached to their original source and version.

Benchmark registry
Artificial Analysis Coding Index - Unversioned - indexTop 8 reported models
GPT-5-Codex
38.9
Step 3.7 Flash
37.1
Step 3.5 Flash 2603
34.6
Gemini 2.5 Pro
32.0
Step 3.5 Flash
31.6
GLM-4.6
29.5
Qwen3 Max
26.4
Mistral Medium 3
13.6
SciCode - Unversioned - percent correctTop 8 reported models
Gemini 2.5 Pro
42.8
GPT-5-Codex
40.9
Step 3.5 Flash
40.4
Step 3.7 Flash
40.0
Gemini 2.5 Flash
39.4
Step 3.5 Flash 2603
38.5
GLM-4.6
38.4
Mistral Medium 3
33.1
Terminal-Bench Hard - Unversioned - success rateTop 8 reported models
GPT-5-Codex
37.9
Step 3.7 Flash
35.6
Step 3.5 Flash 2603
32.6
Step 3.5 Flash
27.3
Gemini 2.5 Pro
26.5
GLM-4.6
25.0
GLM-4.5
22.0
Mistral Medium 3
3.8
Catalog activity

Recent record changes

Detected changes from successful source imports.

Full change log
No recent changes detected

This record has not changed in the retained import history.

Public API

Use this record

Fetch the complete source-linked model record without an API key.

API documentation
Endpoint
GET https://www.modelbench.lol/api/v1/models/mistral/mistral-medium-2505
curl
curl "https://www.modelbench.lol/api/v1/models/mistral/mistral-medium-2505"