GPT-5.2 vs Sarvam 30B

Side-by-side benchmark comparison across agentic, coding, multimodal, knowledge, reasoning, and math workflows.

Agentic
Coding
Multimodal & Grounded
Reasoning
Knowledge
Instruction Following
Multilingual
Mathematics

GPT-5.2· Sarvam 30B

Quick Verdict

Pick GPT-5.2 if you want the stronger benchmark profile. Sarvam 30B only becomes the better choice if you want the cheaper token bill.

GPT-5.2 is clearly ahead on the aggregate, 82 to 48. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.

GPT-5.2's sharpest advantage is in coding, where it averages 70.2 against 34. The single biggest benchmark swing on the page is SWE-bench Verified, 80% to 34%.

GPT-5.2 is also the more expensive model on tokens at $2.00 input / $8.00 output per 1M tokens, versus $0.00 input / $0.00 output per 1M tokens for Sarvam 30B. That is roughly Infinityx on output cost alone. GPT-5.2 gives you the larger context window at 400K, compared with 64K for Sarvam 30B.

Operational tradeoffs

ProviderOpenAISarvam
Price$2.00 / $8.00Free*
Speed73 t/sN/A
TTFT130.34sN/A
Context400K64K

Decision framing

BenchLM keeps the benchmark table and the operator tradeoffs on the same page so a better score does not hide a materially slower, pricier, or smaller-context model.

Runtime metrics show N/A when BenchLM does not have a sourced snapshot for that exact model. The scoring rules and freshness policy are documented on the methodology page.

BenchmarkGPT-5.2Sarvam 30B
AgenticGPT-5.2 wins
Terminal-Bench 2.083%
BrowseComp65.8%35.5%
OSWorld-Verified47.3%
CodingGPT-5.2 wins
SWE-bench Verified80%34%
LiveCodeBench79%
SWE-bench Pro55.6%
HumanEval92.1%
LiveCodeBench v670.0%
Multimodal & Grounded
MMMU86.7%
MMMU-Pro79.5%
OfficeQA Pro95%
RealWorldQA83.3%
OmniDocBench 1.585.7%
Video-MME (with subtitle)86.0%
Video-MME (w/o subtitle)85.8%
MathVision83.0%
We-Math79.0%
DynaMath86.8%
MStar77.1%
SimpleVQA55.8%
ChatCVQA82.1%
CC-OCR70.3%
AI2D_TEST92.2%
CountBench91.9%
ERQA59.8%
VideoMMMU85.9%
MLVU (M-Avg)85.6%
Reasoning
MuSR93%
BBH96%
LongBench v291%
MRCRv293%
ARC-AGI-252.9%
gpqaDiamond66.5%
KnowledgeGPT-5.2 wins
MMLU99%85.1%
GPQA92.4%
SuperGPQA95%
MMLU-Pro88%80%
HLE42%
FrontierScience91%
SimpleQA95%
Instruction Following
IFEval94%
Multilingual
MGSM95%
MMLU-ProX91%
MathematicsGPT-5.2 wins
AIME 202399%
AIME 202499%
AIME 202598%80%
HMMT Feb 202395%
HMMT Feb 202497%
HMMT Feb 202596%
BRUMO 202596%
MATH-50098%97%
HMMT Feb 202573.3%
HMMT Nov 202574.2%
Frequently Asked Questions (5)

Which is better, GPT-5.2 or Sarvam 30B?

GPT-5.2 is ahead overall, 82 to 48. The biggest single separator in this matchup is SWE-bench Verified, where the scores are 80% and 34%.

Which is better for knowledge tasks, GPT-5.2 or Sarvam 30B?

GPT-5.2 has the edge for knowledge tasks in this comparison, averaging 80.2 versus 80. Inside this category, MMLU is the benchmark that creates the most daylight between them.

Which is better for coding, GPT-5.2 or Sarvam 30B?

GPT-5.2 has the edge for coding in this comparison, averaging 70.2 versus 34. Inside this category, SWE-bench Verified is the benchmark that creates the most daylight between them.

Which is better for math, GPT-5.2 or Sarvam 30B?

GPT-5.2 has the edge for math in this comparison, averaging 97.3 versus 86.5. Inside this category, AIME 2025 is the benchmark that creates the most daylight between them.

Which is better for agentic tasks, GPT-5.2 or Sarvam 30B?

GPT-5.2 has the edge for agentic tasks in this comparison, averaging 66.2 versus 35.5. Inside this category, BrowseComp is the benchmark that creates the most daylight between them.

Last updated: April 3, 2026

Weekly LLM Benchmark Digest

Get notified when new models drop, benchmark scores change, or the leaderboard shifts. One email per week.

Free. No spam. Unsubscribe anytime. We only store derived location metadata for consent routing.