{"schemaVersion":"1.1","datasetVersion":"61a32d862da8","capturedAt":"2026-07-22T00:00:00.000Z","model":{"slug":"devstral-small-2505","displayName":"Devstral Small (May '25)","sourceUrl":"https://artificialanalysis.ai/models/devstral-small-2505"},"metrics":{"m1":{"key":"aa-capability-agentic:unversioned:capability_index","benchmark":"aa-capability-agentic","benchmarkName":"Artificial Analysis agentic capability index","version":"unversioned","metric":"capability_index","label":"agentic capability index","unit":"index","higherIsBetter":true,"sourceName":"Artificial Analysis"},"m2":{"key":"aa-capability-coding:unversioned:capability_index","benchmark":"aa-capability-coding","benchmarkName":"Artificial Analysis coding capability index","version":"unversioned","metric":"capability_index","label":"coding capability index","unit":"index","higherIsBetter":true,"sourceName":"Artificial Analysis"},"m3":{"key":"aa-capability-economics:unversioned:capability_index","benchmark":"aa-capability-economics","benchmarkName":"Artificial Analysis economics capability index","version":"unversioned","metric":"capability_index","label":"economics capability index","unit":"index","higherIsBetter":true,"sourceName":"Artificial Analysis"},"m4":{"key":"aa-capability-engineering:unversioned:capability_index","benchmark":"aa-capability-engineering","benchmarkName":"Artificial Analysis engineering capability index","version":"unversioned","metric":"capability_index","label":"engineering capability index","unit":"index","higherIsBetter":true,"sourceName":"Artificial Analysis"},"m5":{"key":"aa-capability-finance:unversioned:capability_index","benchmark":"aa-capability-finance","benchmarkName":"Artificial Analysis finance capability index","version":"unversioned","metric":"capability_index","label":"finance capability index","unit":"index","higherIsBetter":true,"sourceName":"Artificial Analysis"},"m6":{"key":"aa-capability-healthcare:unversioned:capability_index","benchmark":"aa-capability-healthcare","benchmarkName":"Artificial Analysis healthcare capability index","version":"unversioned","metric":"capability_index","label":"healthcare capability index","unit":"index","higherIsBetter":true,"sourceName":"Artificial Analysis"},"m7":{"key":"aa-capability-indices:public-model-v1:agentic_index","benchmark":"aa-capability-indices","benchmarkName":"Artificial Analysis capability indices","version":"public-model-v1","metric":"agentic_index","label":"Agentic Index","unit":"index","higherIsBetter":true,"sourceName":"Artificial Analysis"},"m8":{"key":"aa-capability-indices:public-model-v1:coding_index","benchmark":"aa-capability-indices","benchmarkName":"Artificial Analysis capability indices","version":"public-model-v1","metric":"coding_index","label":"Coding Index","unit":"index","higherIsBetter":true,"sourceName":"Artificial Analysis"},"m9":{"key":"aa-capability-indices:public-model-v1:intelligence_index","benchmark":"aa-capability-indices","benchmarkName":"Artificial Analysis capability indices","version":"public-model-v1","metric":"intelligence_index","label":"Intelligence Index","unit":"index","higherIsBetter":true,"sourceName":"Artificial Analysis"},"m10":{"key":"aa-capability-legal:unversioned:capability_index","benchmark":"aa-capability-legal","benchmarkName":"Artificial Analysis legal capability index","version":"unversioned","metric":"capability_index","label":"legal capability index","unit":"index","higherIsBetter":true,"sourceName":"Artificial Analysis"},"m11":{"key":"aa-capability-strategy:unversioned:capability_index","benchmark":"aa-capability-strategy","benchmarkName":"Artificial Analysis strategy capability index","version":"unversioned","metric":"capability_index","label":"strategy capability index","unit":"index","higherIsBetter":true,"sourceName":"Artificial Analysis"},"m12":{"key":"aa-intelligence-components:4.1:critpt","benchmark":"aa-intelligence-components","benchmarkName":"Artificial Analysis intelligence component evaluations","version":"4.1","metric":"critpt","label":"CritPt","unit":"fraction","higherIsBetter":true,"sourceName":"Artificial Analysis"},"m13":{"key":"aa-intelligence-components:4.1:gdpval_elo","benchmark":"aa-intelligence-components","benchmarkName":"Artificial Analysis intelligence component evaluations","version":"4.1","metric":"gdpval_elo","label":"GDPval-AA","unit":"elo","higherIsBetter":true,"sourceName":"Artificial Analysis"},"m14":{"key":"aa-intelligence-components:4.1:gdpval_normalized","benchmark":"aa-intelligence-components","benchmarkName":"Artificial Analysis intelligence component evaluations","version":"4.1","metric":"gdpval_normalized","label":"GDPval-AA normalized","unit":"fraction","higherIsBetter":true,"sourceName":"Artificial Analysis"},"m15":{"key":"aa-intelligence-components:4.1:gpqa","benchmark":"aa-intelligence-components","benchmarkName":"Artificial Analysis intelligence component evaluations","version":"4.1","metric":"gpqa","label":"GPQA Diamond","unit":"fraction","higherIsBetter":true,"sourceName":"Artificial Analysis"},"m16":{"key":"aa-intelligence-components:4.1:hle","benchmark":"aa-intelligence-components","benchmarkName":"Artificial Analysis intelligence component evaluations","version":"4.1","metric":"hle","label":"Humanity's Last Exam","unit":"fraction","higherIsBetter":true,"sourceName":"Artificial Analysis"},"m17":{"key":"aa-intelligence-components:4.1:lcr","benchmark":"aa-intelligence-components","benchmarkName":"Artificial Analysis intelligence component evaluations","version":"4.1","metric":"lcr","label":"AA-LCR","unit":"fraction","higherIsBetter":true,"sourceName":"Artificial Analysis"},"m18":{"key":"aa-intelligence-components:4.1:omniscience","benchmark":"aa-intelligence-components","benchmarkName":"Artificial Analysis intelligence component evaluations","version":"4.1","metric":"omniscience","label":"AA-Omniscience","unit":"index","higherIsBetter":true,"sourceName":"Artificial Analysis"},"m19":{"key":"aa-intelligence-components:4.1:scicode","benchmark":"aa-intelligence-components","benchmarkName":"Artificial Analysis intelligence component evaluations","version":"4.1","metric":"scicode","label":"SciCode","unit":"fraction","higherIsBetter":true,"sourceName":"Artificial Analysis"},"m20":{"key":"aa-intelligence-components:4.1:tau_banking","benchmark":"aa-intelligence-components","benchmarkName":"Artificial Analysis intelligence component evaluations","version":"4.1","metric":"tau_banking","label":"Tau Banking","unit":"fraction","higherIsBetter":true,"sourceName":"Artificial Analysis"},"m21":{"key":"aa-intelligence-components:4.1:terminalbench_v21","benchmark":"aa-intelligence-components","benchmarkName":"Artificial Analysis intelligence component evaluations","version":"4.1","metric":"terminalbench_v21","label":"Terminal-Bench v2.1","unit":"fraction","higherIsBetter":true,"sourceName":"Artificial Analysis"},"m22":{"key":"aa-model-aggregate:public-model-v1:cache_read_price_per_million_tokens","benchmark":"aa-model-aggregate","benchmarkName":"Artificial Analysis model aggregate","version":"public-model-v1","metric":"cache_read_price_per_million_tokens","label":"缓存读取价格 / 1M Token","unit":"usd_per_1m_tokens","higherIsBetter":false,"sourceName":"Model Coordinate repository snapshots"},"m23":{"key":"aa-model-aggregate:public-model-v1:cache_write_price_per_million_tokens","benchmark":"aa-model-aggregate","benchmarkName":"Artificial Analysis model aggregate","version":"public-model-v1","metric":"cache_write_price_per_million_tokens","label":"缓存写入价格 / 1M Token","unit":"usd_per_1m_tokens","higherIsBetter":false,"sourceName":"Model Coordinate repository snapshots"},"m24":{"key":"aa-model-aggregate:public-model-v1:context_window_tokens","benchmark":"aa-model-aggregate","benchmarkName":"Artificial Analysis model aggregate","version":"public-model-v1","metric":"context_window_tokens","label":"上下文窗口","unit":"tokens","higherIsBetter":true,"sourceName":"Model Coordinate repository snapshots"},"m25":{"key":"aa-model-aggregate:public-model-v1:end_to_end_response_time","benchmark":"aa-model-aggregate","benchmarkName":"Artificial Analysis model aggregate","version":"public-model-v1","metric":"end_to_end_response_time","label":"端到端响应时间","unit":"seconds","higherIsBetter":false,"sourceName":"Model Coordinate repository snapshots"},"m26":{"key":"aa-model-aggregate:public-model-v1:input_price_per_million_tokens","benchmark":"aa-model-aggregate","benchmarkName":"Artificial Analysis model aggregate","version":"public-model-v1","metric":"input_price_per_million_tokens","label":"输入价格 / 1M Token","unit":"usd_per_1m_tokens","higherIsBetter":false,"sourceName":"Model Coordinate repository snapshots"},"m27":{"key":"aa-model-aggregate:public-model-v1:output_price_per_million_tokens","benchmark":"aa-model-aggregate","benchmarkName":"Artificial Analysis model aggregate","version":"public-model-v1","metric":"output_price_per_million_tokens","label":"输出价格 / 1M Token","unit":"usd_per_1m_tokens","higherIsBetter":false,"sourceName":"Model Coordinate repository snapshots"},"m28":{"key":"aa-model-aggregate:public-model-v1:output_tokens_per_second","benchmark":"aa-model-aggregate","benchmarkName":"Artificial Analysis model aggregate","version":"public-model-v1","metric":"output_tokens_per_second","label":"生成速度","unit":"tokens_per_second","higherIsBetter":true,"sourceName":"Model Coordinate repository snapshots"},"m29":{"key":"aa-model-aggregate:public-model-v1:time_to_first_answer_token","benchmark":"aa-model-aggregate","benchmarkName":"Artificial Analysis model aggregate","version":"public-model-v1","metric":"time_to_first_answer_token","label":"首个回答 Token 等待","unit":"seconds","higherIsBetter":false,"sourceName":"Model Coordinate repository snapshots"}},"observations":[{"metric":"m24","observedAt":"2026-07-21T00:00:00.000Z","value":256000},{"metric":"m12","observedAt":"2026-07-21T00:00:00.000Z","value":0},{"metric":"m15","observedAt":"2026-07-21T00:00:00.000Z","value":0.434343434343434},{"metric":"m16","observedAt":"2026-07-21T00:00:00.000Z","value":0.0399257195914578},{"metric":"m9","observedAt":"2026-07-21T00:00:00.000Z","value":11.8246453987468,"provisional":true},{"metric":"m17","observedAt":"2026-07-21T00:00:00.000Z","value":0.266666666666667},{"metric":"m18","observedAt":"2026-07-21T00:00:00.000Z","value":-56.9833333333333},{"metric":"m19","observedAt":"2026-07-21T00:00:00.000Z","value":0.24537037037037}]}