In short
How they differ.
Worked out from the figures below. Prices and specs are the makers' own, read Oct 20, 2026.
- Mistral Medium 3.5 is cheaper per output token: $7.50 against $15 per million tokens, so Kimi K3 costs 2 times as much.
- Kimi K3 reads more at once: a context window of 1M tokens against 256K.
- Both can be downloaded and run on your own machine.
- No independent board we follow ranks both models yet.
Side by side
The specs.
| Kimi K3 | Mistral Medium 3.5 | |
|---|---|---|
| Made by | Moonshot AI · China | Mistral AI · France |
| Released | Not stated | Apr 28, 2026 |
| Price per 1M tokens | $3 in / $15 out | $1.50 in / $7.50 out |
| Context window | 1M | 256K |
| Takes in | text, image, video | text, image |
| Download and run it yourself | Yes (Kimi K3 License) | Yes (Modified MIT) |
| Size | 2.8T parameters, 104B active | 128B (dense) parameters |
| Maker says | “Kimi K3 is an open-weight, native multimodal agentic model and our most capable model to date.” | “Our frontier-class multimodal model optimized for agentic and coding use cases.” |
| Sources | Prices, Oct 20, 2026 · Description | Prices, Oct 20, 2026 · Description |
Independent tests
Where they place.
Position on each board we follow (1 is best). Epoch AI's and Vectara's scores are shown; the other boards are linked for their own figures.
| Board | Kimi K3 | Mistral Medium 3.5 |
|---|---|---|
| Epoch Capabilities IndexEpoch AI, Oct 20, 2026 | #17157.42 | Not listed |
| LMArena textLMArena, Oct 15, 2026 | #18 | Not listed |
| LMArena coding chatsLMArena, Oct 15, 2026 | #12 | Not listed |
| LMArena WebDevLMArena, Oct 19, 2026 | #16 | Not listed |
| LMArena mathLMArena, Oct 15, 2026 | #16 | Not listed |
| Humanity's Last ExamScale AI | #14 | Not listed |
| SWE-Bench ProScale AI, Sep 22, 2026 | #4 | Not listed |
| Vals IndexVals AI, Oct 19, 2026 | #18 | Not listed |
| Vibe Code BenchVals AI, Oct 18, 2026 | #16 | Not listed |
| ProofBenchVals AI, Oct 12, 2026 | #12 | Not listed |
| GPQA Diamond (run by Vals AI, archived)Vals AI, Sep 1, 2026 | #12 | Not listed |
| MMMU Pro (run by Vals AI, archived)Vals AI, Sep 1, 2026 | #11 | Not listed |
More comparisons