Indian Flag
Government Of India
A-
A
A+
/Sarvam AI ORGANISATION
UpvoteUpvote
1

sarvamM

Multilingual, hybrid-reasoning, text-only language model built on Mistral-Small
  • Upvotes1upvotes
  • Downloads76downloads
  • Views1,963views
  • File size0
  • Shares0Shares
  • Last updatedupdated 09 Jul 2025
UpvoteUpvote
1
LICENSE Apache 2.0
SECTOR Sector Agnostic

About Model

sarvam-m is a multilingual, hybrid-reasoning, text-only language model built on Mistral-Small. This post-trained version delivers exceptional improvements over the base model:

  • +20% average improvement on Indian language benchmarks

  • +21.6% enhancement on math benchmarks

  • +17.6% boost on programming benchmarks

Performance gains are even more impressive at the intersection of Indian languages and mathematics, with an outstanding +86% improvement in romanized Indian language GSM-8K benchmarks.

Learn more about sarvam-m in our detailed blog - https://www.sarvam.ai/blogs/sarvam-m

Key Features

  • Hybrid Thinking Mode: A single versatile model supporting both "think" and "non-think" modes. Use the think mode for complex logical reasoning, mathematical problems, and coding tasks, or switch to non-think mode for efficient, general-purpose conversation.

  • Advanced Indic Skills: Specifically post-trained on Indian languages alongside English, embodying a character that authentically reflects and emphasizes Indian cultural values.

  • Superior Reasoning Capabilities: Outperforms most similarly-sized models on coding and math benchmarks, demonstrating exceptional reasoning abilities.

  • Seamless Chatting Experience: Full support for both Indic scripts and romanized versions of Indian languages, providing a smooth and accessible multilingual conversation experience.


Metadata Metadata

Apache 2.0

Sector Agnostic

sarvamai

N.A.

Sarvam AI

Aashay Sachdeva

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

Open

20/08/25 05:45:38

N.A.

0

Tags

  • Transformers
  • safetensors
  • endpoints_compatible
  • Text Generation
  • text-generation-inference
  • conversational
  • gu
  • ta
  • en
  • or
  • bn
  • te
  • pa
  • hi
  • mr
  • ml
  • kn
  • autotrain_compatible
  • base_model:finetune:mistralai/Mistral-Small-3.1-24B-Base-2503
  • base_model:mistralai/Mistral-Small-3.1-24B-Base-2503
  • mistral
  • license:apache-2.0
  • region:us

Related Models Related Models

spk cond tts pflow Hindi
This 150 M-parameter non-autoregressive speech generative model, developed by BharatGen, is designed for speaker-conditioned Text-to-speech in Hindi.
BHARATGEN
  • See Upvoters0
  • Downloads3
  • Views121
  • File Size3.46 GB

spk cond tts pflow Marathi
This 150 M-parameter non-autoregressive speech generative model, developed by BharatGen, is designed for speaker-conditioned Text-to-speech in Marathi.
BHARATGEN
  • See Upvoters0
  • Downloads0
  • Views498
  • File Size3.49 GB

spk cond tts pflow Telugu
This 150 M-parameter non-autoregressive speech generative model, developed by BharatGen, is designed for speaker-conditioned Text-to-speech in Telugu.
BHARATGEN
  • See Upvoters0
  • Downloads0
  • Views84
  • File Size3.38 GB

spk cond tts pflow Bengali
This 150 M-parameter non-autoregressive speech generative model, developed by BharatGen, is designed for speaker-conditioned Text-to-speech in Bengali.
BHARATGEN
  • See Upvoters0
  • Downloads1
  • Views82
  • File Size3.49 GB

spk cond tts pflow Tamil
This 150 M-parameter non-autoregressive speech generative model, developed by BharatGen, is designed for speaker-conditioned Text-to-speech in Tamil.
BHARATGEN
  • See Upvoters0
  • Downloads4
  • Views138
  • File Size3.47 GB

Tamil vb
This 150 M-parameter non-autoregressive speech generative model, developed by BharatGen, is designed for speaker-conditioned Text-to-speech in Tamil.
BHARATGEN
  • See Upvoters0
  • Downloads2
  • Views548
  • File Size3.17 GB

Marathi vb
This 150 M-parameter non-autoregressive speech generative model, developed by BharatGen, is designed for speaker-conditioned Text-to-speech in Marathi.
BHARATGEN
  • See Upvoters0
  • Downloads2
  • Views505
  • File Size3.18 GB

Hindi vb
This 150 M-parameter non-autoregressive speech generative model, developed by BharatGen, is designed for speaker-conditioned Text-to-speech in Hindi.
BHARATGEN
  • See Upvoters0
  • Downloads4
  • Views1,079
  • File Size3.18 GB

Telugu vb
This 150 M-parameter non-autoregressive speech generative model, developed by BharatGen, is designed for speaker-conditioned Text-to-speech in Telugu.
BHARATGEN
  • See Upvoters0
  • Downloads2
  • Views485
  • File Size3.17 GB

Bengali vb
This 150 M-parameter non-autoregressive speech generative model, developed by BharatGen, is designed for speaker-conditioned Text-to-speech in Bengali.
BHARATGEN
  • See Upvoters0
  • Downloads1
  • Views317
  • File Size3.17 GB


Related Models by Organization Related Models by Organization

shuka-v1
Multilingual audio to text model
SARVAM AI
  • See Upvoters3
  • Downloads117
  • Views2,675
  • File Size0

sarvam-1
India's first indic model, pretrained on 4 trillion tokens
SARVAM AI
  • See Upvoters4
  • Downloads301
  • Views6,824
  • File Size0

sarvam-30B
Sarvam-30B is an advanced Mixture-of-Experts (MoE) model with 2.4B non-embedding active parameters, designed primarily for practical deployment. It combines strong reasoning, reliable coding ability, and best-in-class conversational quality across Indian languages. Sarvam-30B is built to run reliably in resource-constrained environments and can handle multilingual voice calls while performing tool calls.
SARVAM AI
  • See Upvoters5
  • Downloads215
  • Views3,713
  • File Size59.92 GB

sarvam-105b
Sarvam-105B is an advanced Mixture-of-Experts (MoE) model with 10.3B active parameters, designed for superior performance across a wide range of complex tasks. It is highly optimized for complex reasoning, with particular strength in agentic tasks, mathematics, and coding.
SARVAM AI
  • See Upvoters12
  • Downloads732
  • Views13,074
  • File Size0

sarvamtranslate
Translation model for 22 Indian Languages
SARVAM AI
  • See Upvoters1
  • Downloads129
  • Views2,506
  • File Size0


Related Models by Sector Related Models by Sector

spk cond tts pflow Hindi
This 150 M-parameter non-autoregressive speech generative model, developed by BharatGen, is designed for speaker-conditioned Text-to-speech in Hindi.
BHARATGEN
  • See Upvoters0
  • Downloads3
  • Views121
  • File Size3.46 GB

spk cond tts pflow Marathi
This 150 M-parameter non-autoregressive speech generative model, developed by BharatGen, is designed for speaker-conditioned Text-to-speech in Marathi.
BHARATGEN
  • See Upvoters0
  • Downloads0
  • Views498
  • File Size3.49 GB

spk cond tts pflow Telugu
This 150 M-parameter non-autoregressive speech generative model, developed by BharatGen, is designed for speaker-conditioned Text-to-speech in Telugu.
BHARATGEN
  • See Upvoters0
  • Downloads0
  • Views84
  • File Size3.38 GB

spk cond tts pflow Bengali
This 150 M-parameter non-autoregressive speech generative model, developed by BharatGen, is designed for speaker-conditioned Text-to-speech in Bengali.
BHARATGEN
  • See Upvoters0
  • Downloads1
  • Views82
  • File Size3.49 GB

spk cond tts pflow Tamil
This 150 M-parameter non-autoregressive speech generative model, developed by BharatGen, is designed for speaker-conditioned Text-to-speech in Tamil.
BHARATGEN
  • See Upvoters0
  • Downloads4
  • Views138
  • File Size3.47 GB

Tamil vb
This 150 M-parameter non-autoregressive speech generative model, developed by BharatGen, is designed for speaker-conditioned Text-to-speech in Tamil.
BHARATGEN
  • See Upvoters0
  • Downloads2
  • Views548
  • File Size3.17 GB

Marathi vb
This 150 M-parameter non-autoregressive speech generative model, developed by BharatGen, is designed for speaker-conditioned Text-to-speech in Marathi.
BHARATGEN
  • See Upvoters0
  • Downloads2
  • Views505
  • File Size3.18 GB

Hindi vb
This 150 M-parameter non-autoregressive speech generative model, developed by BharatGen, is designed for speaker-conditioned Text-to-speech in Hindi.
BHARATGEN
  • See Upvoters0
  • Downloads4
  • Views1,079
  • File Size3.18 GB

Telugu vb
This 150 M-parameter non-autoregressive speech generative model, developed by BharatGen, is designed for speaker-conditioned Text-to-speech in Telugu.
BHARATGEN
  • See Upvoters0
  • Downloads2
  • Views485
  • File Size3.17 GB

Bengali vb
This 150 M-parameter non-autoregressive speech generative model, developed by BharatGen, is designed for speaker-conditioned Text-to-speech in Bengali.
BHARATGEN
  • See Upvoters0
  • Downloads1
  • Views317
  • File Size3.17 GB


Related Models by Task Type Related Models by Task Type

LayoutXLM - Multilingual Document Understanding
A multimodal pre-trained Transformer model designed for multilingual document AI, integrating text, layout, and image data to improve visually rich document understanding across multiple languages.
MICROSOFT CORPORATION (INDIA) PVT. LTD.
  • See Upvoters0
  • Downloads13
  • Views1,023
  • File Size0

LayoutLMv2 - Multimodal Document Understanding
An improved multimodal Transformer model for document AI, integrating text, layout, and image pre-training for enhanced document understanding tasks such as form recognition, document classification, and visual question answering.
MICROSOFT CORPORATION (INDIA) PVT. LTD.
  • See Upvoters0
  • Downloads10
  • Views607
  • File Size0

LayoutLMv3 - Multimodal Document AI
A pre-trained multimodal Transformer model for document AI, integrating unified text and image masking for tasks such as form understanding, receipt processing, and document layout analysis.
MICROSOFT CORPORATION (INDIA) PVT. LTD.
  • See Upvoters0
  • Downloads22
  • Views1,611
  • File Size0

tokenmodel
How people read your model before they download it.
Transformers
AIKOSH OPENSOURCE COMMUNITY
  • See Upvoters0
  • Downloads0
  • Views1
  • File Size0

Gemma-4-12B-it-assistant
This model card is for the Multi-Token Prediction (MTP) drafters for the Gemma 4 models. MTP is implemented by extending the base model with a smaller, faster draft model. When used in a Speculative Decoding pipeline, the draft model predicts several tokens ahead, which the target model then verifies in parallel. This results in significant decoding speedups (up to 3x) while guaranteeing the exact same quality as standard generation, making these checkpoints perfect for low-latency and ODML.
GOOGLE LLC
  • See Upvoters0
  • Downloads3
  • Views32
  • File Size0

DiffusionGemma
DiffusionGemma is a generative model built by Google DeepMind. Based on the 26B A4B Mixture-of-Experts (MoE) Gemma 4 architecture, DiffusionGemma generates tokens using discrete diffusion. This open-weights model is multimodal, handling text, image, and video inputs to generate text output.
GOOGLE LLC
  • See Upvoters1
  • Downloads3
  • Views68
  • File Size0

Med-Doc-Classifier
A document/image classifier for indian healthcare setting
EKA CARE
  • See Upvoters0
  • Downloads1
  • Views135
  • File Size0

Parrotlet-V-2p5-pro
Model for schema-driven extraction from Indian medical records: prescriptions, invoices, lab reports, and discharge summaries, plus PII detection.
EKA CARE
  • See Upvoters0
  • Downloads0
  • Views140
  • File Size0

Document-PII-Redactor
Image PII redaction tuned for indian medical documents
EKA CARE
  • See Upvoters0
  • Downloads4
  • Views169
  • File Size0

NEET-BioBERT
NEET-BioBERT is a fine-tuned version of DistilBERT (base uncased) specifically trained to classify the correct option for NEET-style multiple-choice biology questions. It selects the best answer among four choices (A, B, C, D).
AIKOSH OPENSOURCE COMMUNITY
  • See Upvoters0
  • Downloads8
  • Views165
  • File Size256.10 MB