Model comparison
Magistral 1.0 (Small & Medium) vs Qwen3
- Input
- $0.50 / 1M tokens (Magistral Small); $2.00 / 1M tokens (Magistral Medium)
- Output
- $1.50 / 1M tokens (Magistral Small); $5.00 / 1M tokens (Magistral Medium)
- Context
- 128K
- Released
- 2025-06-10
- Input
- $0.455 / 1M tokens
- Output
- $1.82 / 1M tokens
- Context
- 32K (131K with YaRN)
- Released
- 2025-04-28
Qwen3 scores higher on 3 of the 3 benchmarks both models report; Qwen3 is cheaper on input tokens ($0.5 vs $0.5 per 1M, 1.1×); Magistral 1.0 (Small & Medium) has the larger context window (128K vs 32K).
Specs and pricing side by side
| Spec | Magistral 1.0 (Small & Medium) | Qwen3 |
|---|---|---|
| Maker | Mistral AI | Alibaba (Qwen) |
| Released | 2025-06-10 | 2025-04-28 |
| Context window | 128K | 32K (131K with YaRN) |
| Max output | Not separately published; recommended working… | 32,768 tokens (38,912 for hard reasoning) |
| Input price | $0.50 / 1M tokens (Magistral Small); $2.00 / 1M tokens (Magistral Medium) | $0.455 / 1M tokens |
| Output price | $1.50 / 1M tokens (Magistral Small); $5.00 / 1M tokens (Magistral Medium) | $1.82 / 1M tokens |
| Parameters | Magistral Small: 24B (≈23.6B). Magistral Medium: undisclosed (built on Mistral Medium 3). | 235B total / 22B active (MoE flagship); dense 0.6B-32B and 30B-A3B MoE also released |
| License | Magistral Small: Apache 2.0 (open weights). Magistral Medium: proprietary, API-only. | Apache 2.0 |
| Weights | Open weights | Open weightssame |
| Modalities | Text | Textsame |
| Knowledge cutoff | Not disclosed by Mistral. | not published |
Benchmarks both models report
| Benchmark | Magistral 1.0 (Small & Medium) | Qwen3 |
|---|---|---|
| AIME 2024 | 73.6% | 85.7% |
| AIME 2025 | 64.9% | 81.5% |
| LiveCodeBench v5 | 59.4% | 70.7% |
Scores are as published by each maker or its cited source; harness, reasoning effort and date can differ between rows, so treat small gaps as ties. Sources are listed on each model page.
Which one should you pick?
Pick Magistral 1.0 (Small & Medium) if…
- You need the longer context (128K vs 32K).
- Your use case is: self-hosted, privacy-sensitive reasoning workloads using the open-weight Magistral Small.
Pick Qwen3 if…
- You care about AIME 2024, AIME 2025, LiveCodeBench v5 (it scores higher on 3 shared benchmarks).
- Cost matters: input tokens cost 1.1× less ($0.5 vs $0.5 per 1M).
- Your use case is: self-hosted or private-cloud assistants where open weights and Apache 2.0 licensing are required.
FAQ
Which is better, Magistral 1.0 (Small & Medium) or Qwen3?
Qwen3 scores higher on 3 of the 3 benchmarks both models report; Qwen3 is cheaper on input tokens ($0.5 vs $0.5 per 1M, 1.1×); Magistral 1.0 (Small & Medium) has the larger context window (128K vs 32K). The better pick depends on whether you value benchmark scores, price per token, context length or openness for your workload.
How much cheaper is Qwen3?
Magistral 1.0 (Small & Medium): $0.50 / 1M tokens (Magistral Small); $2.00 / 1M tokens (Magistral Medium) input, $1.50 / 1M tokens (Magistral Small); $5.00 / 1M tokens (Magistral Medium) output. Qwen3: $0.455 / 1M tokens input, $1.82 / 1M tokens output (list API prices as published by each maker).
What are the context windows of Magistral 1.0 (Small & Medium) and Qwen3?
Magistral 1.0 (Small & Medium) offers 128K with up to Not separately published; recommended working window is 40K tokens (performance may degrade beyond that, within the 128K maximum). output; Qwen3 offers 32K (131K with YaRN) with up to 32,768 tokens (38,912 for hard reasoning) output.
Is Magistral 1.0 (Small & Medium) newer than Qwen3?
Magistral 1.0 (Small & Medium) was released 2025-06-10; Qwen3 was released 2025-04-28.