Model
Model explorer

Krutrim-2 (12B)

OPEN
Ola Krutrim · Krutrim family · released Jan 31, 2025

12B Indic LLM built on Mistral-NeMo; 128K context spanning English plus 22 Indian languages.

ReasoningCodingVisionFunction callingTool useAgentic
444.7
Elo · rank #320
Parameters
12B
Active params
12B (dense)
Context
128K tokens
Architecture
Dense Mistral-NeMo-based transformer, 40 layers, GQA (32 heads / 8 KV heads)
License
Krutrim Community License Agreement v1.0
Languages
23+
API price (in/out)
No hosted API
Modalities
text
Benchmark results
Bar shows position within the tracked field; marker = field best
ARC-ChallengeReasoning60.0%#90
best: Llama 3.1 405B · 96.9%
ARC-EasyReasoning82.0%#19
best: Phi-3-medium (14B) · 97.7%
GSM8KMath71.0%#100
best: Llama 3.1 405B · 96.8%
HellaSwagReasoning83.0%#58
best: Claude 3 Opus · 95.4%
HumanEvalCoding80.0%#56
best: Claude Opus 4.5 · 99.4%
IFEvalReasoning73.0%#108
best: Gemma 4 26B A4B · 98.5%
MMLUKnowledge63.0%#192
best: OpenAI o3 · 92.9%
OpenBookQAReasoning49.0%#26
best: Claude 1 · 90.8%
TriviaQAKnowledge62.0%#33
best: Sarvam-1 (2B) · 90.6%
TruthfulQAKnowledge59.0%#26
best: Phi-3.5-MoE (16x3.8B, 6.6B active) · 77.5%
WinoGrandeReasoning77.0%#50
best: PaLM 2 · 90.9%
Run it locally
VRAM @ Q4
7.5 GB
VRAM @ FP16
24.5 GB
Fits on (Q4)
RTX 3060 12GBRTX 4070 Ti 16GBRTX 3090 24GBRTX 4090 24GBRTX 5090 32GBM4 Pro 48GBM3 Max 128GBM3 Ultra 512GBA100 80GBH100 80GBH200 141GBB200 192GB
Throughput data unavailable.
Quantizations
GGUF Q4
Krutrim family
Elo progression across releases
Krutrim-2 (12B)Jan 2025444.7
Resources
HFkrutrim-ai-labs/Krutrim-2-instructGHola-krutrim/Krutrim-2-12BAPIAvailable via Ola Krutrim API
API price weights · each benchmark row carries its own source badge (see methodology)