Skip to content
OSMEXO

Mistral Medium 3.5

mistral/mistral-medium-3.5
  • Mistral
  • Family · Mistral Medium 3
  • Released 29 Apr 2026
  • Text · Image · File → Text
  • Reasoning · adjustable
  • Tools · JSON · Vision
Not on OSMEXO yet
Spec & list price
Context256K
Max output210K
Input / 1M$1.50
Output / 1M$7.505× input
Scores · standing
Intelligence14.2#183/490
Coding46.9#73/196
Usage
Tokens / day392.5M▼ 21% 7d
Usage rank#245
Requests / day60.7K
Tool calls / req0.16
Per request · 30 days
Avg input5.3K0.4× median
Avg output7791.2× median
Cache hit63%about median
Reasoning share50%0.9× median
Tool-call failures1.3%0.7× median

Consumption

833.3M1.7B2.5B4 Oct18 Oct2 Nov
Peak 2.2B 12 OctLow 217.2M 1 NovRank today #245

News & trends

Trending · 14 days
News · latest from Mistral

Benchmarks

bar = share of the top score · Benchmark data: Artificial Analysis
τ²-Bench#29 of 319
94.2%
IFBench#66 of 328
68.8%
Long-context reasoning#149 of 393
69.3%
GPQA Diamond#181 of 445
74.8%
SciCode#106 of 160
40.2%
Coding index#73 of 196
46.9
Terminal-Bench 2.1#76 of 182
50.6%
Terminal-Bench Hard#63 of 311
33.3%
Speed · median across 3 providers, latest day
Output speed#52/331108 tok/s1.7× faster than median
Fastest provider123 tok/sbest of 3 providers
First token#830.62 s1.7× quicker than median
First answer token#4912 sthinking included

#21 cheapest of 26 with a similar score

Value
Intelligence Index
81012141618$0.5$1$2$5Mistral Medium 3.5
Blended price per 1M tokens (log)
Family progress
Intelligence Index
810121416JulSepNov2026MarMistral Medium 3.5
  • Family releases14.2

Technical details

Reasoning
on · high
Tokenizer
Mistral
Moderation
none
Reasoning effort
  • high · default
  • none
Request parameters 14
  • tools
  • tool_choice
  • response_format
  • structured_outputs
  • reasoning
  • reasoning_effort
  • include_reasoning
  • temperature
  • top_p
  • max_tokens
+4 more
  • stop
  • seed
  • frequency_penalty
  • presence_penalty

Interesting facts

  1. 3.2× the output room of similar models (209.7K against a 65.5K median).
  2. Released 188 days ago, the newest of 3 Mistral Medium 3 models.
  3. Weekdays average 795.1M tokens a day against 412.0M on weekends.
  4. 2nd of 3 Mistral Medium 3 models by tokens last month.

When to use

Best for

  • Agentic tool use and multi-step workflows
  • Remote coding agents
  • Strict instruction following in assistants
  • Long document reasoning with image input

Think twice for

  • High-volume simple chat (cheaper options exist)
  • Image generation
  • Self-hosting (closed weights)

Same family

Mistral Medium 3

Alternatives

other vendors · similar capability

Want Mistral Medium 3.5 on OSMEXO?

Sign in to request it. Requests are reviewed; a request does not guarantee availability.