# MGSM / mgsm en > Source: Benchize (https://benchize.com) — https://benchize.com/benchmarks/vals-ai%3Amgsm%3Amgsm_en. When citing or reusing this content, credit "Benchize" and link to https://benchize.com/benchmarks/vals-ai%3Amgsm%3Amgsm_en. A multilingual benchmark for mathematical questions. Canonical URL: https://benchize.com/benchmarks/vals-ai%3Amgsm%3Amgsm_en Last updated: Wed, 21 Oct 2026 18:52:20 GMT Reproduction and quotation must name Benchize, link to the canonical HTML URL, and retain individual source citations. Conquered A multilingual benchmark for mathematical questions. Mathematics · Models tested: 8 · Score publishers: Vals AI ## Model results | # | Model | Score | ± | Cost | Latency | | | --- | --- | --- | --- | --- | --- | --- | | 1 | [MiniMax M2.1](https://benchize.com/models/minimax-m2-1.md) | 100% | SE ±0 pp | $0.00115 | 12.13source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 100 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.001154 \| USD \| per_test \| Lower is better \| \| `latency` \| 12.127 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 131000 provider MiniMax reasoning Not reported Reasoning effort unknown Temperature 1 topP 0.95 verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 75 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mgsm.CneCP-rm.json) Source Vals AI Original model label minimax/MiniMax-M2.1 Source version 1 Observed 2026-10-21T16:46:02.976Z Evidence SHA-256 `1b87f953e52c00bf4bdd90c3add26165f644eeb7a74a8382d2a0324ff0cc1040` Source locator `/tasks/mgsm_en/minimax~1MiniMax-M2.1` | | | | | | | | 2 | [Gemini 3 Flash Preview](https://benchize.com/models/gemini-3-flash-preview-high.md)High | 98.8% | SE ±0.69 pp | $0.00651 | 12.69source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 98.8 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.006511 \| USD \| per_test \| Lower is better \| \| `latency` \| 12.686 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.69 pp sourcePopulation 75 stderr 0.689 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mgsm.CneCP-rm.json) Source Vals AI Original model label google/gemini-3-flash-preview Source version 1 Observed 2026-10-21T16:46:02.976Z Evidence SHA-256 `1b87f953e52c00bf4bdd90c3add26165f644eeb7a74a8382d2a0324ff0cc1040` Source locator `/tasks/mgsm_en/google~1gemini-3-flash-preview` | | | | | | | | 3 | [Mistral Large 3 2512](https://benchize.com/models/mistral-large-2512.md) | 98.4% | SE ±0.79 pp | $0.000383 | 5.29source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 98.4 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.000383 \| USD \| per_test \| Lower is better \| \| `latency` \| 5.293 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 256000 provider Mistral AI reasoning Not reported Reasoning effort unknown Temperature 0.065 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.79 pp sourcePopulation 75 stderr 0.794 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mgsm.CneCP-rm.json) Source Vals AI Original model label mistralai/mistral-large-2512 Source version 1 Observed 2026-10-21T16:46:02.976Z Evidence SHA-256 `1b87f953e52c00bf4bdd90c3add26165f644eeb7a74a8382d2a0324ff0cc1040` Source locator `/tasks/mgsm_en/mistralai~1mistral-large-2512` | | | | | | | | 4 | [Command A](https://benchize.com/models/command-a.md) | 97.6% | SE ±0.97 pp | $0.00238 | 4.65source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 97.6 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.002382 \| USD \| per_test \| Lower is better \| \| `latency` \| 4.648 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 8000 provider Cohere reasoning Not reported Reasoning effort unknown Temperature 0.3 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.97 pp sourcePopulation 75 stderr 0.968 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mgsm.CneCP-rm.json) Source Vals AI Original model label cohere/command-a-03-2025 Source version 1 Observed 2026-10-21T16:46:02.976Z Evidence SHA-256 `1b87f953e52c00bf4bdd90c3add26165f644eeb7a74a8382d2a0324ff0cc1040` Source locator `/tasks/mgsm_en/cohere~1command-a-03-2025` | | | | | | | | 5 | [GPT-4o (2024-08-06)](https://benchize.com/models/gpt-4o-2024-08-06-high.md)High | 97.2% | SE ±1.04 pp | $0.0028 | 6.88source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 97.2 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.0028 \| USD \| per_test \| Lower is better \| \| `latency` \| 6.877 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 16384 provider OpenAI reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.04 pp sourcePopulation 75 stderr 1.043 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mgsm.CneCP-rm.json) Source Vals AI Original model label openai/gpt-4o-2024-08-06 Source version 1 Observed 2026-10-21T16:46:02.976Z Evidence SHA-256 `1b87f953e52c00bf4bdd90c3add26165f644eeb7a74a8382d2a0324ff0cc1040` Source locator `/tasks/mgsm_en/openai~1gpt-4o-2024-08-06` | | | | | | | | 6 | [GPT-4o (2024-11-20)](https://benchize.com/models/gpt-4o-2024-11-20-high.md)High | 96% | SE ±1.24 pp | $0.00357 | 3.56source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 96 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.003573 \| USD \| per_test \| Lower is better \| \| `latency` \| 3.558 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 16384 provider OpenAI reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.24 pp sourcePopulation 75 stderr 1.239 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mgsm.CneCP-rm.json) Source Vals AI Original model label openai/gpt-4o-2024-11-20 Source version 1 Observed 2026-10-21T16:46:02.976Z Evidence SHA-256 `1b87f953e52c00bf4bdd90c3add26165f644eeb7a74a8382d2a0324ff0cc1040` Source locator `/tasks/mgsm_en/openai~1gpt-4o-2024-11-20` | | | | | | | | 7 | [Gemini 2.5 Flash Lite Preview 09-2025](https://benchize.com/models/gemini-2-5-flash-lite-preview-09-2025.md) | 95.6% | SE ±1.3 pp | $0.000146 | 1.1source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 95.6 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.000146 \| USD \| per_test \| Lower is better \| \| `latency` \| 1.1 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65535 provider Google reasoning Not reported Reasoning effort unknown Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.3 pp sourcePopulation 75 stderr 1.297 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mgsm.CneCP-rm.json) Source Vals AI Original model label google/gemini-2.5-flash-lite-preview-09-2025 Source version 1 Observed 2026-10-21T16:46:02.976Z Evidence SHA-256 `1b87f953e52c00bf4bdd90c3add26165f644eeb7a74a8382d2a0324ff0cc1040` Source locator `/tasks/mgsm_en/google~1gemini-2.5-flash-lite-preview-09-2025` | | | | | | | | 8 | [GPT-4o-mini (2024-07-18)](https://benchize.com/models/gpt-4o-mini-2024-07-18-high.md)High | 94.8% | SE ±1.4 pp | $0.000185 | 3.59source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 94.8 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.000185 \| USD \| per_test \| Lower is better \| \| `latency` \| 3.59 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 16384 provider OpenAI reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.4 pp sourcePopulation 75 stderr 1.404 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mgsm.CneCP-rm.json) Source Vals AI Original model label openai/gpt-4o-mini-2024-07-18 Source version 1 Observed 2026-10-21T16:46:02.976Z Evidence SHA-256 `1b87f953e52c00bf4bdd90c3add26165f644eeb7a74a8382d2a0324ff0cc1040` Source locator `/tasks/mgsm_en/openai~1gpt-4o-mini-2024-07-18` | | | | | | | --- Cite as: MGSM / mgsm en — Benchize, https://benchize.com/benchmarks/vals-ai%3Amgsm%3Amgsm_en (retrieved 2026-10-21)