CO
Aya Expanse 32B
Cohere
Mistral AI model product
Mistral Medium 3
Mistral Medium 3 is a multilingual language model from Mistral AI for chat, reasoning, and tool-assisted workflows.
Updated Sep 8, 2026. Default version: Mistral Medium 3
LLMBoard ScoreN/ANot scored
CoverageN/ANot scored
Context window131.1KTokens
Official input price$0.40Mistral API
Mistral Medium 3 Benchmark Results
Benchmark scores for the scored version.
No benchmark results
No benchmark result is linked to the scored version.
Mistral Medium 3 Arena Results
Preference and agent-evaluation results for the default version.
Sort by
Arena | Category | Rank | Rating / score | Votes | Observations | Result date |
|---|
| Arenavision | Categorycreative writing | Rank24 | Rating / score1,168.64 | Votes518 | ObservationsN/A | Result date |
| Arenavision style control | Categorycreative writing | Rank24 | Rating / score1,165.67 | Votes518 | ObservationsN/A | Result date |
| Arenavision | Categorycaptioning | Rank29 | Rating / score1,095.87 | Votes161 | ObservationsN/A | Result date |
| Arenavision style control | Categorycaptioning | Rank29 | Rating / score1,091.78 | Votes161 | ObservationsN/A | Result date |
| Arenavision style control | Categoryentity recognition | Rank41 | Rating / score1,125.53 | Votes199 | ObservationsN/A | Result date |
| Arenavision | Categoryentity recognition | Rank42 | Rating / score1,140.46 | Votes199 | ObservationsN/A | Result date |
| Arenavision | Categoryhumor | Rank65 | Rating / score1,144.20 | Votes442 | ObservationsN/A | Result date |
| Arenavision style control | Categoryhumor | Rank65 | Rating / score1,128.23 | Votes442 | ObservationsN/A | Result date |
| Arenavision | Categorycreative writing vision | Rank69 | Rating / score1,147.96 | Votes480 | ObservationsN/A | Result date |
| Arenavision style control | Categorycreative writing vision | Rank69 | Rating / score1,142.44 | Votes480 | ObservationsN/A | Result date |
| Arenavision style control | Categorychinese | Rank70 | Rating / score1,163.71 | Votes603 | ObservationsN/A | Result date |
| Arenavision | Categorychinese | Rank71 | Rating / score1,156.84 | Votes603 | ObservationsN/A | Result date |
| Arenavision | Categorydiagram | Rank72 | Rating / score1,154.61 | Votes1,428 | ObservationsN/A | Result date |
| Arenavision | Categoryhomework | Rank72 | Rating / score1,165.98 | Votes818 | ObservationsN/A | Result date |
| Arenavision | Categoryocr | Rank72 | Rating / score1,153.73 | Votes4,584 | ObservationsN/A | Result date |
| Arenavision | Categoryenglish | Rank74 | Rating / score1,171.69 | Votes5,329 | ObservationsN/A | Result date |
| Arenavision style control | Categorydiagram | Rank74 | Rating / score1,166.24 | Votes1,428 | ObservationsN/A | Result date |
| Arenavision style control | Categoryhomework | Rank74 | Rating / score1,167.81 | Votes818 | ObservationsN/A | Result date |
| Arenavision style control | Categoryocr | Rank75 | Rating / score1,158.83 | Votes4,584 | ObservationsN/A | Result date |
| Arenavision | Categoryoverall | Rank77 | Rating / score1,156.76 | Votes10,835 | ObservationsN/A | Result date |
| Arenavision style control | Categoryenglish | Rank78 | Rating / score1,164.69 | Votes5,329 | ObservationsN/A | Result date |
| Arenavision style control | Categoryoverall | Rank78 | Rating / score1,155.14 | Votes10,835 | ObservationsN/A | Result date |
| Arenatext style control | Categorygerman | Rank89 | Rating / score1,392.88 | Votes991 | ObservationsN/A | Result date |
| Arenatext style control | Categorypolish | Rank92 | Rating / score1,391.52 | Votes2,876 | ObservationsN/A | Result date |
| Arenatext | Categorypolish | Rank98 | Rating / score1,370.56 | Votes2,876 | ObservationsN/A | Result date |
| Arenatext | Categorygerman | Rank99 | Rating / score1,377.29 | Votes991 | ObservationsN/A | Result date |
| Arenatext | Categoryjapanese | Rank102 | Rating / score1,317.43 | Votes892 | ObservationsN/A | Result date |
| Arenatext style control | Categoryjapanese | Rank104 | Rating / score1,326.11 | Votes892 | ObservationsN/A | Result date |
| Arenatext style control | Categoryindustry legal and government | Rank105 | Rating / score1,409.42 | Votes2,116 | ObservationsN/A | Result date |
| Arenatext style control | Categorymulti turn | Rank108 | Rating / score1,403.28 | Votes5,723 | ObservationsN/A | Result date |
Mistral Medium 3 Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- $0.40 input, $2 output per 1M
- Official provider
- Mistral
- Lowest third-party
- From $0.40 input, $2 output per 1M via NanoGPT
- Tracked offerings
- 8
Sort by
Provider | Provider model ID | Region | Input / 1M | Output / 1M | Context | Updated |
|---|
| ProviderNvidia | Provider model IDmistralai/mistral-medium-3-instruct | Regionglobal | Input / 1MN/A | Output / 1MN/A | Context131.1K | Updated |
| ProviderNanoGPT | Provider model IDmistralai/mistral-medium-3 | Regionglobal | Input / 1M$0.40 | Output / 1M$2 | Context131.1K | Updated |
| ProviderOpenRouter | Provider model IDmistralai/mistral-medium-3 | Regionglobal | Input / 1M$0.40 | Output / 1M$2 | Context131.1K | Updated |
| ProviderMerge Gateway | Provider model IDmistral/mistral-medium-2505 | Regionglobal | Input / 1M$0.40 | Output / 1M$2 | Context128K | Updated |
| ProviderAzure Cognitive Services | Provider model IDmistral-medium-2505 | Regionglobal | Input / 1M$0.40 | Output / 1M$2 | Context128K | Updated |
| ProviderAzure | Provider model IDmistral-medium-2505 | Regionglobal | Input / 1M$0.40 | Output / 1M$2 | Context128K | Updated |
| ProviderMistral | Provider model IDmistral-medium-2505 | Regionglobal | Input / 1M$0.40 | Output / 1M$2 | Context131.1K | Updated |
| ProviderEden AI | Provider model IDmistral/mistral-medium-2505 | Regionglobal | Input / 1M$0.40 | Output / 1M$2 | Context131.1K | Updated |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
Mistral Medium 3 Runtime Performance
Provider-specific output speed and catalog latency for Mistral Medium 3. 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 Medium 3 Specifications
Technical details for the model's default version.
- Version
- Mistral Medium 3
- Released
- May 7, 2025
- Knowledge cutoff
- May 1, 2025
- Parameters
- N/A
- Context window
- 131.1K
- Max output
- 131.1K
- Inputs
- image, text
- Outputs
- text
- Open weights
- No
- License
- Unspecified
Mistral Medium 3 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 Medium 3 | Released | LLMBoardN/A | ParametersN/A | Context131.1K | Max output131.1K | Open weightsNo | LicenseUnspecified |
Models similar to Mistral Medium 3
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
CO
Aya Expanse 8B
Cohere
CO
Aya Vision 32B
Cohere
AN
Claude Fable 5
Anthropic
AN
Claude Fable 5.1
Anthropic
AN
Claude Haiku 3
Anthropic
What is Mistral Medium 3?
Key information about Mistral Medium 3 and its available data.
Mistral Medium 3 is an LLM from Mistral AI. It is for multilingual chat, reasoning, and tool-assisted workflows.
Data as of 2026-09-08.
FAQ
Common questions about Mistral Medium 3.
When was Mistral Medium 3 released?
Mistral Medium 3's default version was released on May 7, 2025.
How much does Mistral Medium 3 cost?
Mistral Medium 3's official API price is $0.40 per million input tokens and $2 per million output tokens via Mistral. The lowest tracked third-party offer starts at $0.40 input and $2 output via NanoGPT.
Who created Mistral Medium 3?
Mistral Medium 3 was created by Mistral AI.
What is the context window for Mistral Medium 3?
The default version has a 131.1K token context window.
Is Mistral Medium 3 open weight?
No. The default version is not marked as having publicly available weights.