Mistral Medium 3.5 vs Step 5 Preview
Mistral AI · France | StepFun · China · Updated June 2026
Quick verdict
Pick Mistral Medium 3.5 for strong intelligence-for-size — artificial analysis intelligence index 30, near the top of its price tier or 128b dense open-weight model — self-hostable. Pick Step 5 Preview for sparse moe architecture — 600b total parameters, only 27b active per token — built for long-horizon agentic work across large codebases and documents or 1m-token context window (64k max output) at $1/$2.70 per million tokens, among the cheapest frontier-scale context windows available. Choose Mistral Medium 3.5 if you need self-hosting or data privacy; Step 5 Preview if you want a managed API.
Mistral Medium 3.5 (Mistral AI, France) and Step 5 Preview (StepFun, China) line up two different AI ecosystems against each other — a comparison that is as much about cost philosophy and openness as raw capability. Mistral Medium 3.5 is mistral's efficient 128B open-weight model — a single dense model unifying reasoning and coding that punches above its size on independent tests (AA Index 30). Step 5 Preview is stepFun's September 20, 2026 flagship preview — a 600B-parameter sparse MoE model (27B active) with a 1M-token context window at $1/$2.70 per million tokens, open weights promised for mid-October. They diverge most on price, context window and open vs. closed weights — each quantified below from the models' real specs.
Key differences at a glance
- ▸Price: Step 5 Preview is about 1.5× cheaper on input ($1/$2.7 per 1M tokens vs $1.5/$7.5 per 1M tokens) — modest, but it adds up at steady volume.
- ▸Context window: Step 5 Preview holds 3.9× more — 1M tokens (~1,500 pages) vs 256K (~384 pages). But effective recall usually fades long before the advertised ceiling, so the bigger number only helps if the model reasons over it.
- ▸Recency: Step 5 Preview is the newer model by about 5 months (released September 20, 2026), usually meaning fresher training data and capabilities.
- ▸Ecosystem: this is a France-vs-China matchup — they differ in pricing philosophy, data-residency options, and tooling ecosystems, not only benchmarks.
Side-by-side specs
| Spec | Mistral Medium 3.5 | Step 5 Preview |
|---|---|---|
| Provider | Mistral AI (France) | StepFun (China) |
| Released | April 28, 2026 | September 20, 2026 |
| Context window | 256K (~384 pages) | 1M tokens (~1,500 pages) |
| Price (in/out) | $1.5/$7.5 per 1M tokens | $1/$2.7 per 1M tokens |
| Open weight? | Yes — self-hostable | No — API only |
| Modalities | text, image, code | text, image, code |
| SWE-Bench Verified | Not published | Not published |
| MRCR v2 @ 1M | Not published | Not published |
Who wins what
Strong intelligence-for-size — Artificial Analysis Intelligence Index 30, near the top of its price tier
Mistral Medium 3.5
Mistral's efficient 128B open-weight model — a single dense model unifying reasoning and coding that punches above its size on independent tests (AA Index 30) — and its weights are open while Step 5 Preview is API-only.
128B dense open-weight model — self-hostable
Mistral Medium 3.5
Open weights make this possible at all — Step 5 Preview is API-only, so it cannot leave the vendor's servers.
Unifies reasoning and coding into one model with an adjustable reasoning effort
Mistral Medium 3.5
Step 5 Preview is comparatively weak here — trails closed frontier models on coding — 67.7% on DeepSWE v1.1 vs GPT-6 Astra's 74.1% and Claude Opus 5's 74.0%
Sparse MoE architecture — 600B total parameters, only 27B active per token — built for long-horizon agentic work across large codebases and documents
Step 5 Preview
Its 1M tokens window holds about 3.9× more than Mistral Medium 3.5's 256K in a single prompt.
1M-token context window (64K max output) at $1/$2.70 per million tokens, among the cheapest frontier-scale context windows available
Step 5 Preview
At $1/$2.7 per 1M tokens it undercuts Mistral Medium 3.5 ($1.5/$7.5 per 1M tokens), and that gap compounds at volume.
67.7% on DeepSWE v1.1, ahead of open-weight rivals Kimi K3 (67.5%) and GLM-5.3 (66.9%)
Step 5 Preview
Mistral Medium 3.5 is comparatively weak here — output pricing ($7.50/M) is higher than the cheapest Chinese rivals
Lowest cost at scale
Step 5 Preview
At $1/$2.7 per 1M tokens, it is the cheaper of the two — the gap dominates the bill on high-volume workloads.
Largest single-prompt input
Step 5 Preview
Its 1M tokens window is about 3.9× larger than Mistral Medium 3.5's 256K, fitting roughly 1,500 pages in one prompt.
Which should you pick?
A cost-sensitive startup shipping high volume
→ Step 5 Preview
At $1/$2.7 per 1M tokens it undercuts Mistral Medium 3.5, and on millions of tokens that margin decides the monthly bill.
Someone analysing very long documents or codebases
→ Step 5 Preview
Larger 1M tokens window fits more in one prompt.
A team with data-privacy or self-hosting needs
→ Mistral Medium 3.5
Open weights let you run it on your own hardware; Step 5 Preview is API-only.
Anyone whose priority is strong intelligence-for-size — artificial analysis intelligence index 30, near the top of its price tier
→ Mistral Medium 3.5
It is specifically built for that.
Anyone whose priority is sparse moe architecture — 600b total parameters, only 27b active per token — built for long-horizon agentic work across large codebases and documents
→ Step 5 Preview
That is its strongest area.
An enterprise with regional data-residency rules
→ Step 5 Preview or Mistral Medium 3.5
Origin (France vs China) affects where data is processed and which compliance regime applies — check the provider's terms for your region.
Mistral Medium 3.5: where it fits
Mistral's efficient 128B open-weight model — a single dense model unifying reasoning and coding that punches above its size on independent tests (AA Index 30). Released April 28, 2026 by Mistral AI, it is built for strong intelligence-for-size — Artificial Analysis Intelligence Index 30, near the top of its price tier, 128B dense open-weight model — self-hostable, unifies reasoning and coding into one model with an adjustable reasoning effort, and 256K context with text and image input.
Its trade-offs are real: below the absolute frontier — a value/efficiency pick, not a flagship-beater, output pricing ($7.50/M) is higher than the cheapest Chinese rivals, license framing ('open' vs enterprise) varies across sources — confirm terms for commercial use, and no native video or audio. At $1.5 in / $7.5 out per million tokens, it sits in the mid price band.
Step 5 Preview: where it fits
StepFun's September 20, 2026 flagship preview — a 600B-parameter sparse MoE model (27B active) with a 1M-token context window at $1/$2.70 per million tokens, open weights promised for mid-October. Released September 20, 2026 by StepFun, it is built for sparse MoE architecture — 600B total parameters, only 27B active per token — built for long-horizon agentic work across large codebases and documents, 1M-token context window (64K max output) at $1/$2.70 per million tokens, among the cheapest frontier-scale context windows available, 67.7% on DeepSWE v1.1, ahead of open-weight rivals Kimi K3 (67.5%) and GLM-5.3 (66.9%), and scores 44 on the Artificial Analysis Intelligence Index, ranking 27th of 653 models tracked.
Its trade-offs: trails closed frontier models on coding — 67.7% on DeepSWE v1.1 vs GPT-6 Astra's 74.1% and Claude Opus 5's 74.0%, open weights were promised for October 15, 2026 but hadn't shipped as of this writing — API-only for now despite the open-weight roadmap, a 'Preview' release, not StepFun's finished flagship — positioning and pricing may change once weights ship, and stepCodeBench and DeepSWE figures are StepFun's own reported numbers, not yet independently reproduced at scale. At $1 in / $2.7 out per million tokens, it sits in the budget price band.
The bottom line for this matchup
The defining split here is open vs. closed. Mistral Medium 3.5 gives you weights you control — self-host it, fine-tune it, keep data in-house, pay only for hardware. Step 5 Preview gives you a managed, always-updated API with no infrastructure to run. Teams with GPUs, privacy requirements, or huge volume often favour the open model; teams that want zero ops and the latest capabilities favour the closed one. Capability is close enough that this operational question, not the benchmark, usually decides it.
Want both Mistral Medium 3.5 and Step 5 Preview without two subscriptions? LumiChats gives you these plus 40+ models under one ₹69/day pass (about $1/day) — draft with one, cross-check with the other.
See pricingFrequently asked questions
Is Mistral Medium 3.5 or Step 5 Preview better for coding?
Public SWE-Bench figures are not available for either model, so the honest test is your own repository — run an identical real bug through both. By design, Mistral Medium 3.5 leans toward strong intelligence-for-size — artificial analysis intelligence index 30, near the top of its price tier while Step 5 Preview leans toward sparse moe architecture — 600b total parameters, only 27b active per token — built for long-horizon agentic work across large codebases and documents, and that positioning usually predicts which feels better on your codebase.
Which is cheaper, Mistral Medium 3.5 or Step 5 Preview?
Mistral Medium 3.5 is open-weight, so self-hosting means no per-token fee (you pay for hardware instead), while Step 5 Preview is API-metered at $1/$2.7 per 1M tokens. For most teams without GPUs, the API model is cheaper to start; at very high volume, self-hosting can win.
Which has the bigger context window?
Step 5 Preview — 1M tokens vs 256K, about 3.9× larger. Useful only if the model actually reasons over the full window, which not all do.
Can I use both Mistral Medium 3.5 and Step 5 Preview together?
Yes — a multi-model platform like LumiChats gives you Mistral Medium 3.5, Step 5 Preview and 40+ others under one ₹69/day pass (about $1/day), so you can draft with one and cross-check with the other instead of buying two subscriptions.
Which is newer, Mistral Medium 3.5 or Step 5 Preview?
Step 5 Preview — released September 20, 2026, about 5 months after Mistral Medium 3.5.
Related comparisons
Specifications and benchmarks reflect publicly reported figures as of June 2026 and may change as providers release updates. Always verify on your own workload.