Mistral Large 3
| Maker | Mistral |
|---|---|
| Family | Mistral Large |
| Released | 2025-12-02 |
| Knowledge cutoff | 2024-11 |
| Context | 262.1K |
| Max output | 209.7K |
|---|---|
| Open weights | Open |
| Modalities | text, image, pdf → text |
| Input / 1M | $0.50 |
| Output / 1M | $1.50 |
| Review call | $0.0167 |
|---|---|
| Hosts | OpenRouter |
| Model id | mistralai/mistral-large-2512 |
CodePanel's benchmark has not run Mistral Large 3 as a pull-request reviewer yet, so this page has no recall, false-alarm or measured-cost figures for it. The methodology explains how ranked models are scored.
No Mistral model is in the run yet. The ranked model closest to it in list price is DeepSeek V4 Flash (Workers AI) from DeepSeek, which caught 49% [37%–60%] of the known bugs at rank 5 of 17, with a measured review cost of $0.0178. At list price, a typical review call (26.2K tokens in and 2.4K out) costs $0.0167 on Mistral Large 3 and $0.0147 on DeepSeek V4 Flash (Workers AI), so this model's call is 14% more than that one's.
A newer Mistral Large model is current: Mistral Large 4.
| Host | Model id | Input / 1M | Output / 1M |
|---|---|---|---|
| OpenRouter | mistralai/mistral-large-2512 | $0.50 | $1.50 |
Source: models.dev's OpenRouter listing, Oct 23, 2026
| Benchmark | Score | Range | Setting | Source | Date |
|---|---|---|---|---|---|
| Arena WebDev | 1231 Arena rating | 1205 to 1256 | – | LMArena CC BY 4.0 | Oct 16, 2026 |
| Version | Released | Input / 1M | Output / 1M | Context |
|---|---|---|---|---|
| Mistral Large 4 | 2026-10-06 | $1.36 | $4.18 | 1M |
| Model | Maker | Input / 1M | Output / 1M | Review call | Context | Bugs caught |
|---|---|---|---|---|---|---|
| Nemotron 3 Super (Workers AI) | NVIDIA | $0.50 | $1.50 | $0.0167 | 256K | – |
| GLM-5.2 | Z.ai | $0.0341 | $7.00 | $0.0178 | 1M | – |
| DeepSeek V4 Flash (Workers AI) | DeepSeek | $0.44 | $1.32 | $0.0147 | 1M | 49% |
| Inkling Small | Thinking Machines | $0.45 | $1.20 | $0.0147 | 524.3K | – |
Updated Oct 23, 2026