Compare / Mercury 3 vs Mistral Large 4
Mercury 3 vs Mistral Large 4
Which is cheaper, which is more capable, and which does better at each kind of work, from 51 results on 4 sources that measured both models.
Of the 51 results both models have, the better value on each. Results whose reported ranges overlap are too close to call.
Results as of 20 October 2026 (catalogue release 2026-10-20-96eb40c3ec63).
Mercury 3 or Mistral Large 4? The short answer
Pick Mercury 3: it scores higher (Spring Prompt overall 39 against 33) and costs less.
- Spring Prompt overall: Mercury 3 39, Mistral Large 4 33 (out of 100; how the score works).
- Mercury 3 is cheaper to run: $0.012 against $0.015 per deck on DeckBench (21% less).
- Mercury 3 is newer: released 19 Oct 2026, against 6 Oct 2026 for Mistral Large 4.
- On our own benchmarks, on the results both models have: Mistral Large 4 does better for marketing planning.
Head to head
Every fact side by side
| Mercury 3 | Mistral Large 4 | |
|---|---|---|
| Price per million tokens | $0.20 in · $0.70 out | $1.36 in · $4.18 out |
| Blended (3 in : 1 out) | $0.33 | $2.06 |
| Cheapest host, blended | – | – |
| Cost per deck (DeckBench) | $0.012 | $0.015 |
| Cached input, per million | $0.020 cached input | – |
| Intelligence Index | – | 38.4 |
| Spring Prompt overall | 39 | 33 |
| Output speed, tokens a second | – | 60 |
| Context window | 1,000,000 tokens | 524,288 tokens |
| Released | 19 Oct 2026 (newer) | 6 Oct 2026 |
| Weights | Closed | Closed |
| Developer based in | the US | France |
Prices are the developer's list price where we have read it, otherwise the typical host on OpenRouter; see every model's price. Intelligence and speed from Artificial Analysis.
Which is better for what
Each model's rank among the models on sale today, for each kind of work, from our own benchmarks and licensed sources. Further right is better; the better rank is in green.
Their slides, side by side
The same brief for both: Manchester store investment review for Hearthside Coffee, an invented company. First slide of each deck from DeckBench.
From our research
- Mercury 3 first look: very fast with thinking off, too slow with it on 20 Oct 2026
- AI Checkpoint: Sol’s new fast lane 9 Oct 2026
- Mistral Large 4 Is Fast, Cheap and Europe's Best Model. On Business Work, It Still Makes Things Up. 7 Oct 2026
- We Told 11 AI Models What Happened in 2026. Most of Them Thought We Were Making It Up. 12 Oct 2026
Every shared result
Each model's best result on every measure both have, grouped by benchmark, with the setting that produced it. Under each measure, every model that benchmark has measured is a grey tick, best to the right, so you can see where the two sit in the field. The headline measures of each benchmark come first; the rest are under "Show all". Green marks the better value; where a source reports ranges that overlap, neither is marked.
| Measure | Mercury 3 | Mistral Large 4 |
|---|---|---|
| BulletBench measured by Spring Prompt · 3 results | ||
| Ladder Elo · Lightning 10+1ladder Elo, higher is better 12 · 18 of 24 measured | 555no reasoning | 249 |
| Cost per game · Lightning 10+1US dollars, lower is better 5 · 20 of 24 measured | $0.0010no reasoning | $0.0117 |
| Games lost on time · Lightning 10+1% of games, lower is better 1 · 18 of 24 measured | 0.0%no reasoning | 58.3% |
| DeckBench measured by Spring Prompt · 3 results | ||
| Cost per deckUS dollars, lower is better 4 · 6 of 24 measured | $0.0116 | $0.0148 |
| Deck ratingrating, higher is betterThe rating weighs each win by how strong the other deck was; the win rate counts wins against whichever decks a model happened to meet, so the two can disagree. 18 · 23 of 24 measured | 812 | 637 |
| Accurate decks% of tasks, higher is better 20 · 20 of 24 measured | 0.0% | 0.0% |
| ROASBench measured by Spring Prompt · 3 results | ||
| Overall scorescore out of 100, higher is better 18 · 14 of 24 measured | 26.1 | 37.0 |
| Audience scorescore out of 100, higher is better 21 · 14 of 24 measured | 43.2 | 55.8 |
| Business scorescore out of 100, higher is better 18 · 14 of 24 measured | 17.6 | 34.3 |
| SurveyBench measured by Spring Prompt · 3 results | ||
| Analysis ratingrating, higher is betterThe rating weighs each win by how strong the other analysis was; the win rate counts wins against whichever analyses a model happened to meet, so the two can disagree. 9 · 7 of 13 measured | 752 | 805 |
| Cost per taskUS dollars, lower is better 3 · 7 of 12 measured | $0.0216 | $0.0924 |
| Analysis scorepoints of 100, higher is better 12 · 8 of 13 measured | 42.2 | 63.8 |
Show all 51 results (39 more)
| Measure | Mercury 3 | Mistral Large 4 |
|---|---|---|
| BulletBench measured by Spring Prompt · 2 results | ||
| Invalid moves · Lightning 10+1% of moves, lower is better 18 · 22 of 24 measured | 0.4%no reasoning | 1.8% |
| Median move time · Lightning 10+1milliseconds, lower is better 3 · 14 of 24 measured | 0.4 sno reasoning | 0.8 s |
| DeckBench measured by Spring Prompt · 14 results | ||
| Caveat dropped% of tasks, lower is better 12 · 12 of 24 measured | 16.7% | 16.7% |
| Clean layout% of tasks, higher is better 6 · 14 of 24 measured | 66.7% | 33.3% |
| Design quality% of the maximum, higher is better 16 · 24 of 24 measured | 58.5% | 30.5% |
| Draft figure quoted% of tasks, lower is better 12 · 9 of 24 measured | 33.3% | 16.7% |
| Findings missing% of tasks, lower is better 17 · 1 of 24 measured | 16.7% | 0.0% |
| Head-to-head win rate% of comparisons, higher is betterThe rating weighs each win by how strong the other deck was; the win rate counts wins against whichever decks a model happened to meet, so the two can disagree. 16 · 23 of 24 measured | 44.0% | 25.0% |
| Layout defects% of tasks, lower is better 6 · 14 of 24 measured | 33.3% | 66.7% |
| Misleading metric used% of tasks, lower is better 22 · 1 of 24 measured | 16.7% | 0.0% |
| Presentable decks% of tasks, higher is better 12 · 12 of 24 measured | 0.0% | 0.0% |
| Recommendation late or wrong% of tasks, lower is better 22 · 22 of 24 measured | 83.3% | 83.3% |
| Slides needing work% of tasks, lower is better 7 · 24 of 24 measured | 33.3% | 100.0% |
| Time per deckseconds, lower is better 9 · 15 of 24 measured | 83.3 | 114 |
| Unsupported claims% of tasks, lower is better 11 · 21 of 24 measured | 16.7% | 50.0% |
| Unsupported numbers% of tasks, lower is better 20 · 23 of 24 measured | 16.7% | 33.3% |
| ROASBench measured by Spring Prompt · 6 results | ||
| Consistency scorescore out of 100, higher is better 19 · 18 of 24 measured | 15.1 | 16.7 |
| Contribution profitsimulated US dollars, higher is better 18 · 12 of 24 measured | $104,868 | $533,401 |
| Cost of a runUS dollars, lower is better 6 · 7 of 24 measured | $0.0533 | $0.0657 |
| Months over budgetmonths of 12, lower is better 1 · 1 of 24 measured | 0 | 0high reasoning |
| Planning scorescore out of 100, higher is better 23 · 4 of 24 measured | 53.8 | 55.0high reasoning |
| Return on ad spendprofit per $1 spent, higher is better 18 · 11 of 24 measured | $0.10 | $0.45 |
| SurveyBench measured by Spring Prompt · 17 results | ||
| Head-to-head win rate% of comparisons, higher is betterThe rating weighs each win by how strong the other analysis was; the win rate counts wins against whichever analyses a model happened to meet, so the two can disagree. 6 · 8 of 13 measured | 48.6% | 35.0% |
| Leading or double-barrelled questions% of tasks, lower is better 8 · 4 of 13 measured | 55.6% | 33.3% |
| Mistaken next steps% of tasks, lower is better 5 · 11 of 13 measured | 22.2% | 66.7% |
| Numbers right% of answers, higher is better 13 · 8 of 13 measured | 63.3% | 98.9% |
| Objectives unanswerable% of tasks, lower is better 1 · 7 of 13 measured | 0.0% | 11.1% |
| Precision ignored% of tasks, lower is better 1 · 8 of 13 measured | 0.0% | 55.6% |
| Research scorepoints of 100, higher is better 11 · 8 of 13 measured | 56.8 | 71.2 |
| Sound research% of tasks, higher is better 6 · 6 of 13 measured | 0.0% | 0.0% |
| Stakeholder's view presumed% of tasks, lower is better 7 · 7 of 13 measured | 22.2% | 22.2% |
| Survey plan scorepoints of 100, higher is better 6 · 7 of 13 measured | 90.8 | 88.4 |
| Survey structure% of tasks, lower is better 7 · 3 of 13 measured | 22.2% | 11.1% |
| Time per taskseconds, lower is better 10 · 12 of 12 measured | 194 | 439 |
| Traps handled% of traps, higher is better 11 · 7 of 13 measured | 60.7% | 78.7% |
| Traps missed% of tasks, lower is better 7 · 10 of 13 measured | 88.9% | 100.0% |
| Unsupported findings% of tasks, lower is better 4 · 10 of 13 measured | 44.4% | 100.0% |
| Wrong numbers% of tasks, lower is better 9 · 7 of 13 measured | 44.4% | 22.2% |
| Wrong recommendation% of tasks, lower is better 12 · 8 of 13 measured | 77.8% | 44.4% |
Compare them with other models
- Mercury 3 vs Claude Opus 5.5Highest-scoring overall available today
- Mercury 3 vs GPT-6 LunaJust above overall
- Mercury 3 vs GLM-5.3Just below overall
- Mistral Large 4 vs Muse Spark 1.3Most intelligent at $1–3 per million tokens
- Mistral Large 4 vs DeepSeek-V4.1-FlashFastest with similar intelligence
- Mistral Large 4 vs Claude Opus 5.5Most intelligent available today
Quick answers
Is Mercury 3 better than Mistral Large 4?
Pick Mercury 3: it scores higher (Spring Prompt overall 39 against 33) and costs less.
Which is cheaper, Mercury 3 or Mistral Large 4?
Mercury 3 is cheaper to run: $0.012 against $0.015 per deck on DeckBench (21% less).
Which is better for decks from an analysis, Mercury 3 or Mistral Large 4?
Mercury 3 ranks 18 of 24 models on sale for turning a finished analysis into a deck you could present as it is, against 23 for Mistral Large 4.
Which is better for user surveys, Mercury 3 or Mistral Large 4?
Mistral Large 4 ranks 7 of 13 models on sale for planning a user survey and reading its results without being misled, against 9 for Mercury 3.
Which is better for marketing planning, Mercury 3 or Mistral Large 4?
Mistral Large 4 ranks 14 of 24 models on sale for planning a year of ad spend without overspending, against 18 for Mercury 3.
Which is better for speed under pressure, Mercury 3 or Mistral Large 4?
Mercury 3 ranks 10 of 27 models on sale for good decisions against a real clock (fast chess), against 17 for Mistral Large 4.
Which has the bigger context window, Mercury 3 or Mistral Large 4?
Mercury 3: 1,000,000 tokens, against 524,288 for Mistral Large 4.
Run Mercury 3 and Mistral Large 4 on your own prompt
Benchmarks aren't your data. Try both side by side in the playground, with the cost of every answer.