Mistral Large 4 wins on price ($0.68 vs $1.5/1M input) and context window (524K vs 256K). Mistral Medium 3.5 wins on coding (91 vs 83). For most workflows, Mistral Large 4 is the stronger default — mistral's open-weight flagship, cheap but still in preview.
The safest Mistral Large 4 vs Mistral Medium 3.5 default, the cheaper option worth trying first, and the specialist pick — before you read the detail below.
Mistral Large 4 is the strongest answer here for Mistral Large 4 vs Mistral Medium 3.5 — pick it when quality of output matters more than the $0.68/1M/1M input you pay for it.
Claude Sonnet 5.5 is the cheaper way in for Mistral Large 4 vs Mistral Medium 3.5, at $2.00/1M/1M input against Mistral Large 4's $0.68/1M/1M.
Mistral Medium 3.5 is the fastest of these for Mistral Large 4 vs Mistral Medium 3.5 — worth it when latency is what the reader notices, not the last few points of reasoning depth.
Mistral Medium 3.5 leads on coding with a score of 91 vs 83 for Mistral Large 4.
Mistral Large 4 has the larger context window: 524K vs 256K for Mistral Medium 3.5.
Mistral Large 4 is cheaper at $0.68/1M input tokens vs $1.5/1M for Mistral Medium 3.5.
Choose Mistral Large 4 for european-hosted and open-weight deployments that need a frontier-class model. Its coding and writing scores are what carry the recommendation here.
Choose Mistral Medium 3.5 when your work is mostly self-hostable European multimodal coding — that is the workload it was tuned for.
Both models serve different primary workflows — Mistral Large 4 for european-hosted and open-weight deployments that need a frontier-class model, Mistral Medium 3.5 for self-hostable European multimodal coding — so running each where it has a clear edge often beats forcing one to do both.
Switch the scoring lens to see whether the Mistral Large 4 vs Mistral Medium 3.5 answer changes when cost, speed, or long-document depth leads the decision.
Mistral / Budget / Oct 10, 2026
Mistral's open-weight flagship, cheap but still in preview.
Ranks models by the broadest mix of coding, writing, research, and long-context usefulness.
You need peak capability today — independent results so far place it mid-table — or you need the weights now.
Where the Mistral Large 4 vs Mistral Medium 3.5 recommendation shifts once you weigh price or latency differently.
Mistral's open-weight flagship, cheap but still in preview.
Best self-hostable multimodal model — European, dense, MIT-licensed.
List prices and published scores — the numbers this page's pick is built from.
| Model | Input | Output | Est. month | Context | Speed | Coding | Writing | Research |
|---|---|---|---|---|---|---|---|---|
| Mistral Large 4Mistral | $0.68/1M | $2.09/1M | $11 | 524k tokens | Balanced | 83 | 85 | 83 |
| Mistral Medium 3.5Mistral | $1.50/1M | $7.50/1M | $30 | 256k tokens | Balanced | 91 | 84 | 82 |
Scores out of 100 — how we evaluate models. “Est. month” is 10M in / 2M out at list price: a ceiling, no discounts.
Why each one is on the shortlist for Mistral Large 4 vs Mistral Medium 3.5, what it is genuinely good at, and where we would steer you away from it.
The default answer for Mistral Large 4 vs Mistral Medium 3.5 — 83/100 on the coding axis, and the model we would start with unless the price below rules it out.
Mistral's October 6, 2026 flagship, in public preview: a natively multimodal mixture-of-experts model with about 1 trillion total parameters, open weights promised for later in October, at $0.68/$2.09 per 1M.
You need peak capability today — independent results so far place it mid-table — or you need the weights now.
Promising, not yet proven. The price is low and the open weights matter for teams that must self-host, but early independent scoring puts it in the middle of the pack. Wait for the weights and more third-party results before building on it.
Full pricing, benchmark table and release notes on the Mistral Large 4 page.
Here for latency: it answers fastest of anything listed for Mistral Large 4 vs Mistral Medium 3.5, at 91/100 on coding.
A 128B dense open-weight multimodal model handling reasoning, coding, and vision in one set of weights — frontier-adjacent coding at mid-tier prices, self-hostable under a modified MIT license.
API price-performance is all that matters — DeepSeek V4-Pro is stronger and cheaper hosted.
The best open-weights model you can realistically self-host. DeepSeek V4 beats it on benchmarks and price via API, but at 128B dense with vision, Medium 3.5 is what you can actually run on your own hardware with EU data residency.
Full pricing, benchmark table and release notes on the Mistral Medium 3.5 page.
UseRightAI recommendations are based on practical decision factors people actually feel in day-to-day use.
Newsletter
We email when the Mistral Large 4 vs Mistral Medium 3.5 pick changes, when one of these models moves on price, or when something new displaces the current leader.
No spam. Useful updates only. Affiliate disclosures always clearly labeled.
Mistral Large 4 wins on more of the categories we score — coding, writing, research — so it is the better default of the two. Mistral Medium 3.5 is the better pick when your work is mostly self-hostable European multimodal coding. Neither is universally "better": Mistral Large 4 is aimed at european-hosted and open-weight deployments that need a frontier-class model, Mistral Medium 3.5 at self-hostable European multimodal coding.
Mistral Large 4 is cheaper at $0.68/1M input and $2.09/1M output. Mistral Medium 3.5 costs $1.5/1M input and $7.5/1M output.
Mistral Large 4 has the larger context window at 524K tokens vs Mistral Medium 3.5's 256K. For large document analysis, Mistral Large 4 is the stronger pick.
Mistral Medium 3.5 is better for coding with a score of 91 vs Mistral Large 4's 83 (out of 100). Claude Opus 5.5 is the overall coding leader in this directory at 100/100.
Both Mistral Large 4 and Mistral Medium 3.5 have similar speed profiles — rated balanced. Neither will be the bottleneck if latency is your deciding factor.
Still a preview: the weights were not downloadable as of October 10, 2026. All benchmark claims so far are Mistral's own; Vals AI ranks it 32nd of 44 on its index. Avoid it if you need peak capability today — independent results so far place it mid-table — or you need the weights now. That is the main case for looking at Mistral Medium 3.5 instead.
Trails GPT-5.6, Opus-class, and Gemini frontier models on complex multi-step reasoning; sparse published benchmark disclosure. 256K context is a quarter of the 1M frontier norm, and $7.50/1M output is dear for the tier. Avoid it if API price-performance is all that matters — DeepSeek V4-Pro is stronger and cheaper hosted. Against Mistral Large 4 specifically, the gap shows up most on coding (83 vs 91).
Take a moderate workload of 10M input and 2M output tokens a month. Mistral Large 4 runs $10.98 (at $0.68/1M in and $2.09/1M out); Mistral Medium 3.5 runs $30.00 (at $1.5/1M in and $7.5/1M out). That is a $19.02/month difference — Mistral Large 4 is the cheaper of the two at this volume, and the gap scales linearly as you send more. Output tokens dominate the bill on both, so prompt length matters far less than response length.
Yes, and for most teams that beats picking one. A common split is Mistral Large 4 for european-hosted and open-weight deployments that need a frontier-class model, with Mistral Medium 3.5 handling self-hostable European multimodal coding. Since Mistral Large 4 is both the stronger and the cheaper option here, a split mainly makes sense if Mistral Medium 3.5 covers a capability you specifically need.