MAI-Thinking-1 vs Mistral Large 4.
Verdict by category, tale of the tape, shared benchmarks only, and cost at three workloads — all from the tracked data.
Verdict by category
| Category | MAI-Thinking-1 | Mistral Large 4 | Winner |
|---|---|---|---|
| League Score | 48.7 | 53.1 | Mistral Large 4 |
| Coding (SWE-bench %) | — | — | — |
| Reasoning (GPQA %) | 84.2 | — | — |
| Context (tokens) | 1000000 | 1000000 | Draw |
| Speed (tok/s) | — | — | — |
Tale of the tape
| Spec | MAI-Thinking-1 | Mistral Large 4 |
|---|---|---|
| Maker | Microsoft AI | Mistral AI |
| Released | Jun 2, 2026 | Oct 6, 2026 |
| League Score | 49 | 53 |
| Context | 1,000,000 tokens | 1,000,000 tokens |
| Max output | — | — |
| $/1M input | — | $0.70 |
| $/1M output | — | $2.10 |
| Weights | Closed | Open |
| Speed | — | — |
Head-to-head benchmarks (shared only)
| Benchmark | MAI-Thinking-1 | Mistral Large 4 |
|---|---|---|
| No shared benchmark measurements yet — shown as not measured, never interpolated. | ||
Cost at three workloads
| Workload | MAI-Thinking-1 | Mistral Large 4 |
|---|---|---|
| Light — 1M in / 250K out | — | $1.23 |
| Standard — 10M in / 2.5M out | — | $12.25 |
| Long-context — 50M in / 5M out | — | $45.50 |
The pick
Pick Mistral Large 4 if you want the higher overall League Score; you want downloadable weights / self-hosting.
All figures from the model database — sources and dates on each model page. Cost your own workload or compare a different pair.
Frequently asked questions
Which is better overall, MAI-Thinking-1 or Mistral Large 4?
Mistral Large 4 leads on overall League Score, 53 to 49 — a gap of 4 points on a composite of sourced benchmarks. That is overall form, not every category: check the per-category verdicts before choosing.
Which is cheaper, MAI-Thinking-1 or Mistral Large 4?
Only Mistral Large 4 has published API pricing in our database ($1.05 blended per 1M tokens); MAI-Thinking-1 has no tracked API price, so a like-for-like cost verdict is not possible.
Which has the bigger context window, MAI-Thinking-1 or Mistral Large 4?
Neither — both take 1M of context.
Are MAI-Thinking-1 and Mistral Large 4 open source?
MAI-Thinking-1 is closed weights — API access only; Mistral Large 4 is open weights — it can be downloaded and self-hosted.