The AI arena is free today

Open Superagent

DeepSeek-V3.2-Speciale vs Sarvam-105B

DeepSeek-V3.2-Speciale leads the LLM Stats Score 34.3 to 25.9.

DeepSeek · Sarvam AI · Updated for 2026

Which is better?

DeepSeek-V3.2-Speciale leads the overall LLM Stats Score 34.3 to 25.9, ranking #87 overall.

In the 4 individual benchmarks reported for both models, DeepSeek-V3.2-Speciale wins 3; this is a narrower head-to-head signal than the composite indexes.

Based on current LLM Stats indexes, shared benchmarks, pricing, and model metadata for 2026.

Choose DeepSeek-V3.2-Speciale

  • overall performance matters — it scores 34.3 and ranks #87 on LLM Stats
  • your work emphasizes reasoning and coding — it leads those capability indexes
  • you value its reported benchmark strengths — it wins 3 of 4 exact shared results

Choose Sarvam-105B

  • you want the most recent training data — it shipped Mar 2026

At a glance

The differences that matter most.

Core performance indexes
34.3
#87
25.9
#144
32.9
#95
26.4
#138
18.4
#105
-1.3
#238
10.4
#105
6.2
#133
Cost, coverage & limits
Benchmark wins
3 of 4
1 of 4
Input price
$0.28 / M
— / M
Output price
$0.42 / M
— / M
Context window
131,072

Capability indexes

Additional strengths measured across groups of related public benchmarks

1 shared
Index
DeepSeek-V3.2-Speciale
Sarvam-105B
35.1#43
29.2#84
Conservative TrueSkill rating · higher is betterHow scores work

Individual benchmarks

8 reported for DeepSeek-V3.2-Speciale · 14 for Sarvam-105B

4 shared

DeepSeek-V3.2-Speciale outperforms in 3 benchmarks (HMMT 2025, Humanity's Last Exam, SWE-Bench Verified), while Sarvam-105B is better at 1 benchmark (AIME 2025).

DeepSeek-V3.2-Speciale shows notably better performance in the majority of benchmarks.

Sun Sep 06 2026 • llm-stats.com

Human preference

Blind head-to-head votes and playground preference scores

Model Size

Parameter count comparison

580.0B diff

DeepSeek-V3.2-Speciale has 580.0B more parameters than Sarvam-105B, making it 552.4% larger.

DeepSeek
DeepSeek-V3.2-Speciale
685.0Bparameters
Sarvam AI
Sarvam-105B
105.0Bparameters
685.0B
DeepSeek-V3.2-Speciale
105.0B
Sarvam-105B

Context Window

Maximum input and output token capacity

Only DeepSeek-V3.2-Speciale specifies input context (131,072 tokens). Only DeepSeek-V3.2-Speciale specifies output context (131,072 tokens).

DeepSeek
DeepSeek-V3.2-Speciale
Input131,072 tokens
Output131,072 tokens
Sarvam AI
Sarvam-105B
Input- tokens
Output- tokens
Sun Sep 06 2026 • llm-stats.com

License

Usage and distribution terms

DeepSeek-V3.2-Speciale is licensed under MIT, while Sarvam-105B uses Apache 2.0.

License differences may affect how you can use these models in commercial or open-source projects.

DeepSeek-V3.2-Speciale

MIT

Open weights

Sarvam-105B

Apache 2.0

Open weights

Release Timeline

When each model was launched

DeepSeek-V3.2-Speciale was released on 2025-12-01, while Sarvam-105B was released on 2026-03-06.

Sarvam-105B is 3 months newer than DeepSeek-V3.2-Speciale.

DeepSeek-V3.2-Speciale

Dec 1, 2025

9 months ago

Sarvam-105B

Mar 6, 2026

6 months ago

3mo newer

Knowledge Cutoff

When training data ends

Neither model specifies a knowledge cutoff date.

Unable to compare the recency of their training data.

No cutoff dates available

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion

Judge for yourself.

Run your own prompts against DeepSeek-V3.2-Speciale and Sarvam-105B side-by-side, then vote on the output you prefer.

DeepSeek-V3.2-Speciale
✓ Preferred
Sarvam-105B
Open in Playground

FAQ

Common questions about DeepSeek-V3.2-Speciale vs Sarvam-105B.

Which is better, DeepSeek-V3.2-Speciale or Sarvam-105B?

DeepSeek-V3.2-Speciale leads the LLM Stats Score 34.3 to 25.9. DeepSeek-V3.2-Speciale is made by DeepSeek and Sarvam-105B is made by Sarvam AI. The best choice depends on your use case — compare their capability indexes, individual benchmarks, pricing, and limits above.

How does DeepSeek-V3.2-Speciale compare to Sarvam-105B in benchmarks?

DeepSeek-V3.2-Speciale scores HMMT 2025: 99.2%, AIME 2025: 96.0%, CodeForces: 90.0%, t2-bench: 80.3%, SWE-Bench Verified: 73.1%. Sarvam-105B scores MATH-500: 98.6%, AIME 2025: 96.7%, MMLU: 90.6%, HMMT 2025: 85.8%, HMMT25: 85.8%.

What are the context window sizes for DeepSeek-V3.2-Speciale and Sarvam-105B?

DeepSeek-V3.2-Speciale supports 131K tokens and Sarvam-105B supports an unknown number of tokens. A larger context window lets you process longer documents, conversations, or codebases in a single request.

What are the main differences between DeepSeek-V3.2-Speciale and Sarvam-105B?

Key differences include LLM Stats Score (34.3 vs 25.9), licensing (MIT vs Apache 2.0). See the full comparison above for benchmark-by-benchmark results.

Who makes DeepSeek-V3.2-Speciale and Sarvam-105B?

DeepSeek-V3.2-Speciale is developed by DeepSeek and Sarvam-105B is developed by Sarvam AI.