Mistral AI model product
Mistral Medium 3.5
5 is Mistral AI's first flagship merged model: a dense 128B-parameter multimodal model with a 256k context window that unifies instruction-following, reasoning, and coding capabilities in a single set of weights.
Updated Aug 17, 2026. Default version: Mistral Medium 3.5
LLMBoard Score48.7Mistral Medium 3.5
Coverage60%8 benchmark families
Context window262.1KTokens
Official input price$1.5Mistral API
Mistral Medium 3.5 Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
Mistral Medium 3.5 Benchmark Results
Benchmark scores for Mistral Medium 3.5.
Sort by
Benchmark | Score | Rank | Participants | Percentile | Evidence | Evaluated |
|---|
| BenchmarkTau3 Airline | Score72.0% | Rank01 | Participants1 | Percentile100.0% | EvidenceC | Evaluated |
| BenchmarkTau3 Retail | Score76.1% | Rank01 | Participants1 | Percentile100.0% | EvidenceC | Evaluated |
| BenchmarkTau3 Telecom | Score91.4% | Rank01 | Participants1 | Percentile100.0% | EvidenceC | Evaluated |
| BenchmarkBeyond AIME | Score66.9% | Rank04 | Participants5 | Percentile25.0% | EvidenceC | Evaluated |
| BenchmarkCOLLIE | Score95.8% | Rank04 | Participants10 | Percentile66.7% | EvidenceC | Evaluated |
| BenchmarkTau3 Banking | Score13.4% | Rank05 | Participants7 | Percentile33.3% | EvidenceC | Evaluated |
| BenchmarkFinance Agent v2 | Score32.1% | Rank22 | Participants26 | Percentile16.0% | EvidenceB | Evaluated |
| BenchmarkIFBench | Score69.0% | Rank24 | Participants34 | Percentile30.3% | EvidenceC | Evaluated |
| BenchmarkSWE-Bench Verified | Score77.6% | Rank26 | Participants111 | Percentile77.3% | EvidenceC | Evaluated |
| BenchmarkBrowseComp | Score48.6% | Rank49 | Participants62 | Percentile21.3% | EvidenceC | Evaluated |
| BenchmarkAIME 2025 | Score86.3% | Rank59 | Participants115 | Percentile49.1% | EvidenceC | Evaluated |
Mistral Medium 3.5 Arena Results
Preference and agent-evaluation results for the default version.
Sort by
Arena | Category | Rank | Rating / score | Votes | Observations | Result date |
|---|
| Arenaagent bash recovery steps | Categoryoverall | Rank37 | Rating / score-0.0 | VotesN/A | Observations5.9K | Result date |
| Arenaagent | Categoryoverall | Rank40 | Rating / score-0.1 | VotesN/A | Observations168.1K | Result date |
| Arenaagent praise complaint | Categoryoverall | Rank40 | Rating / score-0.1 | VotesN/A | Observations1.1K | Result date |
| Arenaagent task outcome explicit | Categoryoverall | Rank44 | Rating / score-0.1 | VotesN/A | Observations3.2K | Result date |
| Arenaagent steerability | Categoryoverall | Rank45 | Rating / score-0.1 | VotesN/A | Observations4.1K | Result date |
| Arenaagent tool hallucination | Categoryoverall | Rank47 | Rating / score-0.0 | VotesN/A | Observations153.8K | Result date |
| Arenavision | Categoryoverall | Rank64 | Rating / score1221.2 | Votes5,110 | ObservationsN/A | Result date |
| Arenavision style control | Categoryoverall | Rank73 | Rating / score1198.2 | Votes5,110 | ObservationsN/A | Result date |
| Arenatext | Categoryoverall | Rank85 | Rating / score1425.0 | Votes94,069 | ObservationsN/A | Result date |
| Arenavision | Categoryoverall | Rank87 | Rating / score1172.1 | Votes41,623 | ObservationsN/A | Result date |
| Arenatext | Categoryoverall | Rank93 | Rating / score1420.9 | Votes11,022 | ObservationsN/A | Result date |
| Arenavision style control | Categoryoverall | Rank93 | Rating / score1158.7 | Votes41,623 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryoverall | Rank94 | Rating / score1430.7 | Votes10,963 | ObservationsN/A | Result date |
| Arenatext style control | Categoryoverall | Rank95 | Rating / score1427.0 | Votes11,022 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryoverall | Rank97 | Rating / score1429.1 | Votes66,982 | ObservationsN/A | Result date |
| Arenawebdev | Categoryoverall | Rank99 | Rating / score1265.2 | Votes2,154 | ObservationsN/A | Result date |
| Arenatext style control | Categoryoverall | Rank130 | Rating / score1409.1 | Votes94,069 | ObservationsN/A | Result date |
Mistral Medium 3.5 Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- $1.5 input, $7.5 output per 1M
- Official provider
- Mistral
- Lowest third-party
- From $1.5 input, $7.5 output per 1M via Eden AI
- Tracked offerings
- 2
Sort by
Provider | Provider model ID | Region | Input / 1M | Output / 1M | Context | Updated |
|---|
| ProviderMistral | Provider model IDmistral-medium-2604 | Regionglobal | Input / 1M$1.5 | Output / 1M$7.5 | Context262.1K | Updated |
| ProviderEden AI | Provider model IDmistral/mistral-medium-2604 | Regionglobal | Input / 1M$1.5 | Output / 1M$7.5 | Context262.1K | 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 3.5 Runtime Performance
Provider-specific output speed and catalog latency for Mistral Medium 3.5. Runtime does not affect the capability score.
Sort by
Provider | Output Speed | Catalog Latency | Max Input | Max Output | Updated |
|---|
| ProviderMistral AI | Output Speed0.961 tok/s | Catalog Latency10.401 s | Max Input256K | Max Output256K | Updated |
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
Mistral Medium 3.5 Specifications
Technical details for the model's default version.
- Version
- Mistral Medium 3.5
- Released
- Apr 29, 2026
- Knowledge cutoff
- Unknown
- Parameters
- 128B
- Context window
- 262.1K
- Max output
- 262.1K
- Inputs
- image, text
- Outputs
- text
- Open weights
- Yes
- License
- Modified MIT License
Mistral Medium 3.5 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 3.5 | Released | LLMBoard48.7 | Parameters128B | Context262.1K | Max output262.1K | Open weightsYes | LicenseModified MIT License |
Mistral Medium 3.5 vs nearby models
Open a comparison with the three ranked models immediately above and below this model.
Models similar to Mistral Medium 3.5
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
MA
Mistral Small 4
Mistral AI
MA
Ministral 3
Mistral AI
MA
Devstral Medium
Mistral AI
MA
Pixtral Large
Mistral AI
MA
Min istral 3
Mistral AI
MA
Magistral Small
Mistral AI
What is Mistral Medium 3.5?
Key information about Mistral Medium 3.5 and its available data.
5 is Mistral AI's first flagship merged model: a dense 128B-parameter multimodal model with a 256k context window that unifies instruction-following, reasoning, and coding capabilities in a single set of weights. 1 and Magistral in Le Chat, and Devstral 2 in the Vibe coding agent.
Reasoning effort is configurable per request, and the vision encoder was trained from scratch to handle variable image sizes and aspect ratios. Released as open weights under a Modified MIT License.
Data as of 2026-08-17.
FAQ
Common questions about Mistral Medium 3.5.
When was Mistral Medium 3.5 released?
Mistral Medium 3.5's default version was released on Apr 29, 2026.
How much does Mistral Medium 3.5 cost?
Mistral Medium 3.5's official API price is $1.5 per million input tokens and $7.5 per million output tokens via Mistral. The lowest tracked third-party offer starts at $1.5 input and $7.5 output via Eden AI.
Who created Mistral Medium 3.5?
Mistral Medium 3.5 was created by Mistral AI.
What is the context window for Mistral Medium 3.5?
The default version has a 262.1K token context window.
Is Mistral Medium 3.5 open weight?
Yes. The default version is marked as open weight under Modified MIT License.
How many API providers offer Mistral Medium 3.5?
2 provider offerings are linked to the default version.
What models should I compare Mistral Medium 3.5 with?
Nearby ranked alternatives include GPT-5, GLM 4.5, GPT-5.4-nano.