Very long inputs
Choose Grok 4.20 Reasoning
Grok 4.20 Reasoning reads up to 1M tokens per request, against 256k for Ministral 3 8B.
Grok 4.20 Reasoning vs Ministral 3 8B: price, context and capabilities
Choose Grok 4.20 Reasoning for very long inputs and the latest release. Choose Ministral 3 8B for self-hosting and data control.
In short
- A price comparison is not possible: one of them has no comparable published price.
- Grok 4.20 Reasoning reads more per request: 1M tokens against 256k.
- Grok 4.20 Reasoning is the newer model (March 2026, against December 2025).
- Only Grok 4.20 Reasoning offers reasoning, structured output, prompt caching, batch api.
- Only Ministral 3 8B offers self-hostable.
Which should you choose?
Each answer below comes straight from the published prices, specifications and independent benchmark results.
Self-hosting and data control
Choose Ministral 3 8B
Ministral 3 8B has open weights, so you can run it on your own servers; Grok 4.20 Reasoning is only available through its provider.
Latest release
Choose Grok 4.20 Reasoning
Grok 4.20 Reasoning is the newer release (March 2026, against December 2025 for Ministral 3 8B).
Specifications side by side
| Grok 4.20 Reasoning | Ministral 3 8B | |
|---|---|---|
| Provider | xAI | Mistral AI |
| Category | Language models | Language models |
| Status | Generally available | Generally available |
| Headline price | $1.56 / 1M tokens | No data |
| Input | $1.25 / 1M tokens | No data |
| Output | $2.50 / 1M tokens | No data |
| Cached input | $0.20 / 1M tokens | No data |
| Context window | 1M | 256k |
| Max output | No data | No data |
| Released | March 2026 | December 2025 |
| Knowledge cutoff | No data | No data |
| Licence | Proprietary | Open weights |
| Capabilities | Vision, Reasoning, Tool calling, Structured output, Prompt caching, Batch API | Vision, Tool calling, Self-hostable |
| API model id | grok-4.20-0309-reasoning | ministral-8b-2512 |
Benchmarks side by side
Leader in bold.
| Benchmark | Grok 4.20 Reasoning | Ministral 3 8B |
|---|---|---|
| Intelligence Index | No data | 5.5 |
| Terminal-Bench Hard | No data | 4.5% |
| SciCode | No data | 20.7% |
| GPQA Diamond | No data | 47.1% |
| Humanity's Last Exam | No data | 4.3% |
| IFBench (instruction following) | No data | 29.1% |
| τ²-Bench (agentic tool use) | No data | 26.6% |
| MMMU-Pro (visual reasoning) | No data | 46% |
| Long-context reasoning (AA-LCR) | No data | 25.7% |
Independent evaluations by Artificial Analysis, run the same way for every model, at the reasoning setting shown. Percentages unless marked as an index.
Frequently asked questions
Should I choose Grok 4.20 Reasoning or Ministral 3 8B?
Choose Grok 4.20 Reasoning for very long inputs and the latest release. Choose Ministral 3 8B for self-hosting and data control.
Which is cheaper, Grok 4.20 Reasoning or Ministral 3 8B?
A price comparison is not possible: one of them has no comparable published price.
Which has the larger context window, Grok 4.20 Reasoning or Ministral 3 8B?
Grok 4.20 Reasoning reads more per request: 1M tokens against 256k.
What is the difference between Grok 4.20 Reasoning and Ministral 3 8B?
Grok 4.20 Reasoning is made by xAI and Ministral 3 8B by Mistral AI. A price comparison is not possible: one of them has no comparable published price. Grok 4.20 Reasoning reads more per request: 1M tokens against 256k. Grok 4.20 Reasoning is the newer model (March 2026, against December 2025). Only Grok 4.20 Reasoning offers reasoning, structured output, prompt caching, batch api. Only Ministral 3 8B offers self-hostable.
Which is newer, Grok 4.20 Reasoning or Ministral 3 8B?
Grok 4.20 Reasoning was released on March 2026 and Ministral 3 8B on December 2025.