Mistral AI, released Mar 17, 2025
Magistral Medium (latest)price, context, benchmarks and release details
Provisional: not enough results to rank yet 16% confidence 16 percent, Low confidence, 1 of 7 expected sources in
44.1
SI Score
Not ranked yet
- Input, per 1M tokens
- $1.50LiteLLMFirst-party API Standard token rate; MIT LiteLLM transcription. Provider documentation: https://docs.mistral.ai/models/model-cards/mistral-medium-3-5-26-04. Exact endpoint only; cache/batch/long-context rates excluded
Retrieved Oct 9, 2026 · MIT
Open source ↗ - Output, per 1M tokens
- $7.50LiteLLMFirst-party API Standard token rate; MIT LiteLLM transcription. Provider documentation: https://docs.mistral.ai/models/model-cards/mistral-medium-3-5-26-04. Exact endpoint only; cache/batch/long-context rates excluded
Retrieved Oct 9, 2026 · MIT
Open source ↗ - Context window
- 128Kmodels.devPublished source fact
Retrieved Oct 9, 2026 · MIT
Open source ↗ - Max output
- 16.4Kmodels.devPublished source fact
Retrieved Oct 9, 2026 · MIT
Open source ↗ - Released
- Mar 17, 2025models.devPublished source fact
Retrieved Oct 9, 2026 · MIT
Open source ↗
How the score breaks down
Coding (weight 40 percent) —
Math (weight 15 percent) —
Preference (weight 15 percent) —
Reasoning (weight 30 percent) 3.6
Weights: reasoning 30%, math 15%, coding 40%, preference 15%. Results use fixed 0–100 scales before averaging, and thin evidence is pulled toward 50. Method si-v3-retained-evidence-2, computed Oct 9, 2026, 06:15 UTC.
Around it on the leaderboard
- 1 Claude Fable 5.1 80.2
- 2 Claude Opus 5.5 78.4
- 3 GPT-6 Astra 77.5
- 4 Claude Fable 5 76.8
- 5 Claude Opus 5 74.9
Benchmark results
4 benchmarks, 8 resultsEach row shows the best published result. Where a model was tested at several settings, such as reasoning effort, open the row to see each one. Hover or tap a value for its source.
ARC-AGI-1 (public eval)reasoning 8.9%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] magistral-medium-2506Published Oct 6, 2026
Retrieved Oct 9, 2026 · factual citation
Open source ↗
8.9 2 settings
- Setting 1 8.9%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] magistral-medium-2506Published Oct 6, 2026
Retrieved Oct 9, 2026 · factual citation
Open source ↗ - Setting 2 8.0%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] magistral-medium-2506-thinkingPublished Oct 6, 2026
Retrieved Oct 9, 2026 · factual citation
Open source ↗
ARC-AGI-1 (semi-private)reasoning 6.1%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] magistral-medium-2506-thinkingPublished Oct 6, 2026
Retrieved Oct 9, 2026 · factual citation
Open source ↗
6.1 2 settings
- Setting 1 5.9%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] magistral-medium-2506Published Oct 6, 2026
Retrieved Oct 9, 2026 · factual citation
Open source ↗ - Setting 2 6.1%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] magistral-medium-2506-thinkingPublished Oct 6, 2026
Retrieved Oct 9, 2026 · factual citation
Open source ↗
ARC-AGI-2 (public eval)reasoning 0%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] magistral-medium-2506Published Oct 6, 2026
Retrieved Oct 9, 2026 · factual citation
Open source ↗
0.0 2 settings
- Setting 1 0%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] magistral-medium-2506Published Oct 6, 2026
Retrieved Oct 9, 2026 · factual citation
Open source ↗ - Setting 2 0%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] magistral-medium-2506-thinkingPublished Oct 6, 2026
Retrieved Oct 9, 2026 · factual citation
Open source ↗
ARC-AGI-2 (semi-private)reasoning 0%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] magistral-medium-2506Published Oct 6, 2026
Retrieved Oct 9, 2026 · factual citation
Open source ↗
0.0 2 settings
- Setting 1 0%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] magistral-medium-2506Published Oct 6, 2026
Retrieved Oct 9, 2026 · factual citation
Open source ↗ - Setting 2 0%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. [variant] magistral-medium-2506-thinkingPublished Oct 6, 2026
Retrieved Oct 9, 2026 · factual citation
Open source ↗
Normalization uses a fixed 0–100 scale for each unit, independent of other models. Compare evaluation conditions before reading a small gap as decisive. “Lab-reported” marks the provider's own published figure.
Details and sources
- Open weights
- Nomodels.devPublished source fact
Retrieved Oct 9, 2026 · MIT
Open source ↗ - License
- not yet reported
- Input modalities
- textmodels.devPublished source fact
Retrieved Oct 9, 2026 · MIT
Open source ↗ - First seen by SuperIndex
- Oct 8, 2026
- Coverage
- 13% of expected source weight
Reported (1)
- ARC PrizeOct 8, 2026
Awaiting (6)
- Epoch AI Benchmarking25% of weight
- Humanity’s Last Exam13% of weight
- Official model cards via models.dev4% of weight
- LiveBench13% of weight
- LMArena / Arena25% of weight
- Terminal-Bench6% of weight
Confidence rises as pending sources publish. Some sources never cover some models, so confidence reaches 100% at 80% of expected weight.