Mistral Large 2 vs Kimi K3
| Mistral Large 2 | Kimi K3 | |
|---|---|---|
| Fournisseur | Mistral | Moonshot |
| Fenêtre de contexte | 128 000 | 1 000 000 |
| Date de sortie | 2024-07-24 | 2026-05-20 |
| Prix d'entrée | $2.00/ 1M tokens | — |
| Prix de sortie | $6.00/ 1M tokens | — |
| Idéal pour | coding · reasoning · chatbot · multilingual | — |
Capacités
| Mistral Large 2 | Kimi K3 | |
|---|---|---|
| Vision | ✗ | ✗ |
| Raisonnement | ✓ | ✗ |
| Codage | ✓ | ✗ |
| Audio | ✗ | ✗ |
| Appel d'outils | ✓ | ✗ |
| Contexte long | ✓ | ✗ |
Comparaison de classement
| Mistral Large 2 | Kimi K3 | |
|---|---|---|
| Global | 73.6 | 6 |
| Benchmark | 88 | 0 |
| Capacité | 57.1 | 0 |
| Efficacité prix | 100 | 0 |
| Contexte | 64 | 100 |
Global = Benchmark 50% + Capacité 20% + Efficacité prix 20% + Contexte 10%.
Comparaison de benchmarks
| Mistral Large 2 | Kimi K3 | |
|---|---|---|
| Codage | — | — |
| Raisonnement | — | — |
| Mathématiques | — | — |
| Vision | — | — |
Historique des prix
Mistral Large 2
| Date | Entrée | Sortie |
|---|---|---|
| 2026-08-11 | $2.00 | $6.00 |
/ 1M tokens
Kimi K3
Aucun historique de prix pour l'instant.
/ 1M tokens
Mistral Large 2
Mistral's flagship model with strong reasoning and code generation in multiple languages.
Kimi K3
Kimi K3 flagship model
Mistral Large 2
Data Reliability
Benchmark · VerifiedOfficial · Experimental
Data confidence: 90%
Kimi K3
Data Reliability
Benchmark · ExperimentalOfficial · Experimental
Data confidence: 50%
Why compare these models?
Choose Mistral Large 2 if:
- • Higher overall score
Choose Kimi K3 if:
- • General purpose workloads
Frequently Asked Questions
Which model is better: Mistral Large 2 or Kimi K3?
Based on overall scores, Mistral Large 2 ranks higher (73.6 vs 6). See the comparison table for a full breakdown.
Which model is cheaper: Mistral Large 2 or Kimi K3?
Price data is not available for Mistral Large 2 or Kimi K3.
What are the capability differences between Mistral Large 2 and Kimi K3?
Mistral Large 2: Raisonnement, Codage, Appel d'outils, Contexte long