RankingEffort benchmarks
Mistral Large 2
10 measured results · 1 effort settings · 10 benchmark/harness combinations
Change model
Capability profile
One configuration, seven capabilities
- AMistral Large 2 · Non-reasoning3 of 7 capabilities supported
Gaps are unknown, not zero. Hollow points are preliminary.
Agentic
A3.4PreliminaryHard reasoning
A1.6SupportedCoding
A1.9SupportedHuman pref
UnknownKnowledge
A1.9SupportedMultimodal
UnknownLong context
Unknown
Each profile belongs to the named configuration. Capability scores use a shared panel for all axes, distinct from the overall ranking panel; they are not accuracy percentages or direct head-to-head wins.
Capability score evidence
Mistral Large 2 · Non-reasoning: 0.3 overall comparison score. Meets the overall evidence requirements.
Which sources carry the weight?
- Artificial Analysis · 100.0% direct comparison weight · independent of the model provider
Which benchmark families carry the weight?
- MMLU-Pro · 10.6% direct comparison weight
- LiveCodeBench · 10.5% direct comparison weight
- math500 · 10.5% direct comparison weight
- Aime · 10.4% direct comparison weight
- IFBench · 10.3% direct comparison weight
- Other families (5) · 47.7% direct comparison weight
These shares describe direct comparison weight, not fractions of the final score or confidence. Source balancing does not give every family equal weight or remove correlations between overlapping sources.
How sensitive is this score?
The audit removes one source at a time and refits against the same overall reference panel. These published results use the matching dataset audited 2026-11-22.
Both ranks use the same surviving configurations for each removal. The cohort starts with 602 configurations supported in the original audit, including separate effort settings. These are diagnostic positions, not the default leaderboard ranks. Eligibility is not rechecked after removal.
| Removed source | Rank before → after | Rank movement | Cohort remaining / lost support | Score after (change) |
|---|---|---|---|---|
| Artificial Analysis | Comparison support lost; no rank | 230 remaining / 372 lost | Unavailable | |
| ARC Prize | 533 → 531 | Up 2 | 602 remaining / 0 lost | 0.4 |
| Datacurve | 533 → 533 | No movement | 602 remaining / 0 lost | 0.3 |
| LiveCodeBench | 533 → 533 | No movement | 602 remaining / 0 lost | 0.3 |
| Scale | 533 → 533 | No movement | 602 remaining / 0 lost | 0.4 |
| Terminal-Bench | 533 → 533 | No movement | 602 remaining / 0 lost | 0.3 |
| Cognition | 533 → 533 | No movement | 602 remaining / 0 lost | 0.3 |
| LiveBench | 532 → 532 | No movement | 601 remaining / 1 lost | 0.4 |
| SWE-rebench | 533 → 533 | No movement | 602 remaining / 0 lost | 0.3 |
| Epoch AI | 519 → 518 | Up 1 | 588 remaining / 14 lost | 0.3 |
| Vals AI | 532 → 532 | No movement | 601 remaining / 1 lost | 0.3 |
| Provider reports · Anthropic | 533 → 533 | No movement | 602 remaining / 0 lost | 0.3 |
| Provider reports · Moonshot AI | 533 → 533 | No movement | 602 remaining / 0 lost | 0.3 |
| Provider reports · Meta | 533 → 533 | No movement | 602 remaining / 0 lost | 0.3 |
| Provider reports · Mistral | 533 → 533 | No movement | 602 remaining / 0 lost | 0.3 |
| Provider reports · OpenAI | 533 → 533 | No movement | 602 remaining / 0 lost | 0.3 |
| All provider reports | 533 → 533 | No movement | 602 remaining / 0 lost | 0.3 |
Lost support means a configuration no longer has a comparison path to the complete reference panel. Its rank is unavailable. A surviving path alone does not establish leaderboard eligibility.
Overlapping sources may remain correlated even with source balancing. Leave-one-source-out movement is a sensitivity diagnostic, not a spread of source-specific ranks or a confidence measure. Scoring methodology · Dated configuration audit
Every benchmark family
| Removed evidence | Recomputed score | Change |
|---|---|---|
| arc-agi | 0.4 | +0.1 |
| deepswe | 0.3 | -0.1 |
| critpt | 0.3 | -0.1 |
| tau-bench | 0.3 | -0.1 |
| hle | 0.4 | +0.0 |
| math500 | 0.4 | +0.0 |
| aime | 0.4 | +0.0 |
| mmmu-pro | 0.4 | +0.0 |
| terminal-bench | 0.3 | -0.0 |
| epoch-game-puzzles | 0.3 | -0.0 |
| simpleqa | 0.4 | +0.0 |
| mmlu-pro | 0.4 | +0.0 |
| aa-lcr | 0.3 | -0.0 |
| frontiercode | 0.3 | -0.0 |
| livecodebench | 0.3 | +0.0 |
| swe-rebench | 0.3 | +0.0 |
| livebench-instructions | 0.3 | +0.0 |
| enigma-eval | 0.3 | +0.0 |
| multi-swe-bench | 0.3 | -0.0 |
| arc-agi-3 | 0.3 | -0.0 |
| apex-agents | 0.3 | +0.0 |
| livebench-language | 0.3 | +0.0 |
| gdp-pdf | 0.3 | +0.0 |
| enterprise-ops | 0.3 | -0.0 |
| livebench-coding | 0.3 | -0.0 |
| gmmlu | 0.3 | +0.0 |
| terminal-bench-science | 0.3 | +0.0 |
| gdpval | 0.3 | +0.0 |
| swe-atlas-qna | 0.3 | -0.0 |
| vibe-code | 0.3 | -0.0 |
| ifbench | 0.3 | -0.0 |
| gpqa | 0.3 | -0.0 |
| analyst-agent | 0.3 | +0.0 |
| automationbench | 0.3 | -0.0 |
| briefcase | 0.3 | -0.0 |
| livebench-reasoning | 0.3 | -0.0 |
| vals-code-migration | 0.3 | -0.0 |
| itbench | 0.3 | -0.0 |
| vals-legal-research | 0.3 | -0.0 |
| vals-finance-agent | 0.3 | -0.0 |
| vals-tax-agent | 0.3 | -0.0 |
| swe-bench-pro | 0.3 | +0.0 |
| vals-excel-modeling | 0.3 | -0.0 |
| swe-atlas-test-writing | 0.3 | -0.0 |
| harvey | 0.3 | +0.0 |
| omniscience | 0.3 | +0.0 |
| frontiermath | 0.3 | -0.0 |
| livebench-data | 0.3 | +0.0 |
| swe-atlas-refactoring | 0.3 | -0.0 |
| livebench-math | 0.3 | +0.0 |
What evidence is missing from the fit?
Across all collected settings for this model: 10 observations, 10 contributing, 0 matched without graph weight and 0 excluded. Counts do not establish rank eligibility.
Benchmark results
Reviewed retained observations are in use. October 2026 Real-SWE board has no 95% confidence whisker, so these scores stay at the 2026-10-18 retrieval. Original source retrieval 2026-10-18; source updated 2026-10-18.
| Benchmark / source | Non-reasoning |
|---|---|
AIME 2025Benchmark detailsIndependently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-11-22. | 14% Reported settingsMistral Large 2 (Nov '24); Reported effort: non-reasoning; Model source: https://artificialanalysis.ai/models/mistral-large-2; Mode from evaluator isReasoning metadata; effort budget not specified |
CritPtBenchmark detailsBenchmark developed by Argonne and UIUC, with contributions from 60+ researchers globally · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-11-22. | 0% Reported settingsMistral Large 2 (Nov '24); Reported effort: non-reasoning; Model source: https://artificialanalysis.ai/models/mistral-large-2; Mode from evaluator isReasoning metadata; effort budget not specified |
GPQA DiamondBenchmark detailsIndependently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-11-22. | 48.59% Reported settingsMistral Large 2 (Nov '24); Reported effort: non-reasoning; Model source: https://artificialanalysis.ai/models/mistral-large-2; Mode from evaluator isReasoning metadata; effort budget not specified |
Humanity's Last ExamBenchmark detailsIndependently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-11-22. | 3.34% Reported settingsMistral Large 2 (Nov '24); Reported effort: non-reasoning; Model source: https://artificialanalysis.ai/models/mistral-large-2; Mode from evaluator isReasoning metadata; effort budget not specified |
IFBenchBenchmark detailsIndependently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-11-22. | 31.22% Reported settingsMistral Large 2 (Nov '24); Reported effort: non-reasoning; Model source: https://artificialanalysis.ai/models/mistral-large-2; Mode from evaluator isReasoning metadata; effort budget not specified |
LiveCodeBenchBenchmark detailsIndependently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-11-22. | 29.31% Reported settingsMistral Large 2 (Nov '24); Reported effort: non-reasoning; Model source: https://artificialanalysis.ai/models/mistral-large-2; Mode from evaluator isReasoning metadata; effort budget not specified |
MATH-500Benchmark detailsIndependently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-11-22. | 73.6% Reported settingsMistral Large 2 (Nov '24); Reported effort: non-reasoning; Model source: https://artificialanalysis.ai/models/mistral-large-2; Mode from evaluator isReasoning metadata; effort budget not specified |
MMLU-ProBenchmark detailsIndependently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-11-22. | 69.7% Reported settingsMistral Large 2 (Nov '24); Reported effort: non-reasoning; Model source: https://artificialanalysis.ai/models/mistral-large-2; Mode from evaluator isReasoning metadata; effort budget not specified |
𝜏²-Bench TelecomBenchmark detailsIndependently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-11-22. | 30.7% Reported settingsMistral Large 2 (Nov '24); Reported effort: non-reasoning; Model source: https://artificialanalysis.ai/models/mistral-large-2; Mode from evaluator isReasoning metadata; effort budget not specified |
Terminal-Bench HardBenchmark detailsIndependently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-11-22. | 6.06% Reported settingsMistral Large 2 (Nov '24); Reported effort: non-reasoning; Model source: https://artificialanalysis.ai/models/mistral-large-2; Mode from evaluator isReasoning metadata; effort budget not specified |
What these comparisons mean
Each source reports its own effort labels. Matching labels do not establish equal compute budgets or identical fallback behavior. Open Reported settings to inspect the exact model name, agent and task coverage. Multiple results at one level remain visible rather than selecting the best. Source-reported None or Non-reasoning labels are not treated as evidence that an API supports those settings.
These measurements use their own dated source collection. The leaderboard fits identified configurations jointly; it contains no pooled model ratings. Explicit thinking budgets are separate configurations. Read the methodology.
Download all effort benchmark results · 10,184 results across 381 models · Collected 2026-11-22