Sarvam M (Reasoning)

May 23, 2025Last synced May 28, 2026, 9:09 AM

Human Frontier

—

Human-calibrated frontier signal, backed by Arena-style preference evidence and separate from raw AgMoBench benchmark composite scores.

Blended Price

Free/M

Input Price

Free/M

Output Price

Free/M

Speed

141 tok/s

TTFT

1.17s

Benchmark Scores

Intelligence Index

Performance Radar

Sarvam M (Reasoning)

Field Average

External Benchmarks

AA-Omniscience Accuracy

How Sarvam M (Reasoning) Compares

Axes

X Axis

Y Axis

Bubble Size

Filters

Blended Price (USD)$0.00 – $37.5

AgMoBench Overall≥ 2.1

Providers

ai2ai21-labsalibabaanthropicarceeawsazurebaidu

Show quadrants

mistral

meta

openai

google

anthropic

azure

nvidia

alibaba

aws

ibm

nous-research

kimi

xai

zai

deepcogito

kwaikat

xiaomi

cohere

ai21-labs

inclusionai

minimax

baidu

deepseek

prime-intellect

liquidai

stepfun

bytedance_seed

upstage

reka-ai

inception

tencent

arcee

ai2

swiss-ai-initiative

perplexity

Bubble size = Context Window

8191.00

2000000.00

Compare with other models

GPQA Diamond

0.4 / 100

LiveCodeBench

0.3 / 100

Terminal-Bench Hard

0.0 / 100

τ²-Bench

0.0 / 100

AA-Omniscience Hallucination RatePredicted

93.4 / 100

Aider PolyglotPredicted

11.1 / 100

AIME 2025Predicted

0.1 / 30

AlpacaEval 2.0 LCPredicted

40.7 / 100

AlpacaEval 2.0 RawPredicted

32.1 / 100

ARC-AGI-1Predicted

1.0 / 100

ARC-AGI-1 Cost per TaskPredicted

0.0

ARC-AGI-2Predicted

0.0 / 100

ARC-AGI-2 Cost per TaskPredicted

0.0

BFCL (Berkeley Function Calling)Predicted

30.0

BigCodeBench CompletePredicted

57.5 / 100

BigCodeBench InstructPredicted

47.6 / 100

AA Intelligence Index (Matrix)Predicted

66.8

AA Long Context Reasoning (Matrix)Predicted

68.8

AIME 2024Predicted

40.9

AIME 2025 (Matrix)Predicted

8.1

Arena-Hard AutoPredicted

47.6

BrowseCompPredicted

54.4

BRUMO 2025Predicted

57.1

CMIMC 2025Predicted

58.0

GPQA Diamond (Matrix)Predicted

43.8

HLE (Matrix)Predicted

8.2

HMMT Feb 2025Predicted

20.3

HMMT Nov 2025Predicted

87.1

HumanEvalPredicted

76.1

IFBench (Matrix)Predicted

24.3

IFEvalPredicted

76.3

IMO 2025Predicted

7.2

LiveCodeBench (Matrix)Predicted

34.2

MATH-500 (Matrix)Predicted

88.1

MathArena Apex 2025Predicted

0.4

MMLU-Pro (Matrix)Predicted

68.4

MMMU-ProPredicted

80.0

MRCR v2Predicted

72.5

OSWorldPredicted

31.9

SimpleQAPredicted

14.2

SMT 2025Predicted

70.1

SWE-bench ProPredicted

27.6

Tau-Bench Telecom (Matrix)Predicted

93.6

Terminal-Bench 2.0Predicted

12.6

Terminal-Bench 1.0Predicted

9.5

USAMO 2025Predicted

3.3

Video-MMUPredicted

86.0

browsecompPredicted

51.0

BullshitBenchPredicted

8.3 / 100

Aider PolyglotPredicted

0.0

Apex AgentsPredicted

2.0

Arc Agi 2Predicted

0.0

BIG-Bench HardPredicted

3.0

CAD-EvalPredicted

8.3

Chess PuzzlesPredicted

0.1

CyBenchPredicted

0.1

DeepResearchBenchPredicted

0.3

FictionLiveBenchPredicted

0.3

GeoBenchPredicted

0.0

GSM8K (Epoch)Predicted

18.3

HellaSwagPredicted

67.3

Lech Mazur WritingPredicted

6.6

METR Time HorizonsPredicted

0.1

OTIS Mock AIME 2024–2025Predicted

0.1

PosttrainbenchPredicted

0.0

SimpleQA Verified (Epoch)Predicted

0.2

The Agent CompanyPredicted

1.0

TriviaQAPredicted

1.7

WinoGrandePredicted

0.8

FrontierMathPredicted

8.7 / 100

GAIA Level 1Predicted

20.7

GAIA Level 2Predicted

6.4

GAIA Level 3Predicted

0.0

GAIAPredicted

12.8 / 100

LegalBenchPredicted

83.0 / 100

LiveBench CodingPredicted

61.1 / 100

LiveBench Data AnalysisPredicted

33.2 / 100

LiveBench LanguagePredicted

41.0 / 100

LiveBench MathPredicted

49.0 / 100

LiveBench OverallPredicted

34.7 / 100

LiveBench ReasoningPredicted

25.4 / 100

LongBench v2 EasyPredicted

38.7

LongBench v2 HardPredicted

38.5

LongBench v2Predicted

34.8 / 100

MathVistaPredicted

54.5 / 100

MedQA (USMLE)Predicted

69.8

MLE-benchPredicted

10.4 / 100

MMMUPredicted

57.0 / 100

MMTU Table UnderstandingPredicted

43.6 / 100

MT-BenchPredicted

7.9 / 10

NoLiMa (NIAH)Predicted

91.4 / 100

OCRBench v2Predicted

77.6 / 100

Open LLM AveragePredicted

48.2 / 100

Open LLM: BBHPredicted

70.8 / 100

Open LLM: GPQAPredicted

36.2 / 100

Open LLM: IFEvalPredicted

80.9 / 100

Open LLM: MATH Level 5Predicted

55.2 / 100

Open LLM: MMLU-PROPredicted

57.4 / 100

Open LLM: MUSRPredicted

45.6 / 100

RE-BenchPredicted

0.0

SimpleBenchPredicted

14.7 / 100

simpleqaPredicted

13.9

SWE-bench LitePredicted

8.6 / 100

SWE-bench VerifiedPredicted

23.5 / 100

tau-bench RetailPredicted

70.8 / 100

Vectara Factual ConsistencyPredicted

94.8 / 100

Vectara Hallucination RatePredicted

5.2 / 100

WebArenaPredicted

10.0 / 100

WeirdMLPredicted

14.0 / 100

WildBenchPredicted

53.0