Gemini 3 Flash vs Sarvam 105B

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

Agentic
Coding
Multimodal & Grounded
Reasoning
Knowledge
Instruction Following
Multilingual
Mathematics

Gemini 3 Flash· Sarvam 105B

Quick Verdict

Pick Gemini 3 Flash if you want the stronger benchmark profile. Sarvam 105B only becomes the better choice if knowledge is the priority or you want the cheaper token bill.

Gemini 3 Flash is clearly ahead on the aggregate, 67 to 60. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.

Gemini 3 Flash's sharpest advantage is in agentic, where it averages 57.5 against 49.5. The single biggest benchmark swing on the page is MMLU, 70% to 90.6%. Sarvam 105B does hit back in knowledge, so the answer changes if that is the part of the workload you care about most.

Gemini 3 Flash is also the more expensive model on tokens at $0.50 input / $3.00 output per 1M tokens, versus $0.00 input / $0.00 output per 1M tokens for Sarvam 105B. That is roughly Infinityx on output cost alone. Sarvam 105B is the reasoning model in the pair, while Gemini 3 Flash is not. That usually helps on harder chain-of-thought-heavy tests, but it can also mean more latency and more token spend in real use. Gemini 3 Flash gives you the larger context window at 1M, compared with 128K for Sarvam 105B.

Operational tradeoffs

ProviderGoogleSarvam
Price$0.50 / $3.00Free*
Speed159 t/sN/A
TTFT1.19sN/A
Context1M128K

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.

BenchmarkGemini 3 FlashSarvam 105B
AgenticGemini 3 Flash wins
Terminal-Bench 2.056%
BrowseComp66%49.5%
OSWorld-Verified53%
Claw-Eval47.1%
CodingTie
HumanEval62%
SWE-bench Verified44%45%
LiveCodeBench36%
SWE-bench Pro44%
SWE-Rebench52.5%
LiveCodeBench v671.7%
Multimodal & Grounded
MMMU-Pro80%
OfficeQA Pro79%
Reasoning
MuSR65%
BBH84%
LongBench v275%
MRCRv276%
gpqaDiamond78.7%
hle11.2%
KnowledgeSarvam 105B wins
MMLU70%90.6%
GPQA69%
SuperGPQA67%
MMLU-Pro72%81.7%
HLE6%
FrontierScience65%
SimpleQA67%
Instruction FollowingGemini 3 Flash wins
IFEval85%84.8%
Multilingual
MGSM85%
MMLU-ProX78%
MathematicsSarvam 105B wins
AIME 202370%
AIME 202472%
AIME 202571%88.3%
HMMT Feb 202366%
HMMT Feb 202468%
HMMT Feb 202567%
BRUMO 202569%
MATH-50080%98.6%
HMMT Feb 202585.8%
HMMT Nov 202585.8%
Frequently Asked Questions (6)

Which is better, Gemini 3 Flash or Sarvam 105B?

Gemini 3 Flash is ahead overall, 67 to 60. The biggest single separator in this matchup is MMLU, where the scores are 70% and 90.6%.

Which is better for knowledge tasks, Gemini 3 Flash or Sarvam 105B?

Sarvam 105B has the edge for knowledge tasks in this comparison, averaging 81.7 versus 54. Inside this category, MMLU is the benchmark that creates the most daylight between them.

Which is better for coding, Gemini 3 Flash or Sarvam 105B?

Gemini 3 Flash and Sarvam 105B are effectively tied for coding here, both landing at 45 on average.

Which is better for math, Gemini 3 Flash or Sarvam 105B?

Sarvam 105B has the edge for math in this comparison, averaging 92.3 versus 72.6. Inside this category, MATH-500 is the benchmark that creates the most daylight between them.

Which is better for agentic tasks, Gemini 3 Flash or Sarvam 105B?

Gemini 3 Flash has the edge for agentic tasks in this comparison, averaging 57.5 versus 49.5. Inside this category, BrowseComp is the benchmark that creates the most daylight between them.

Which is better for instruction following, Gemini 3 Flash or Sarvam 105B?

Gemini 3 Flash has the edge for instruction following in this comparison, averaging 85 versus 84.8. Inside this category, IFEval 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.