Skip to main content

Model comparison

Exaone 4.0 32B vs Sarvam 30B

Data verified

Head-to-head evidence from 11 shared benchmark results across 5 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.

LG AI Research
40.44/100
Margin
0.3pts
winning →
40.7/100
0 category wins0 category wins

Public leaderboard positions: Exaone 4.0 32B #170 (Estimated); Sarvam 30B #169 (Estimated). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.

Evidence parity. Exaone 4.0 32B and Sarvam 30B share 11 comparable benchmark results. 0 of 8 categories are comparable. 2 results are unique to Exaone 4.0 32B; 0 to Sarvam 30B.

Updated July 23, 2026
Shared results
11
Exaone 4.0 32B only
2
Sarvam 30B only
0
Comparable categories
0 / 8

Benchmark data for Exaone 4.0 32B and Sarvam 30B is coming soon on BenchLM.

Confidence note. This is a partial-evidence comparison with 11 shared benchmark results across 5 evidence categories; 0 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.

Why this result

BenchLM has partial data for these models, but not enough overlapping benchmark coverage to produce a fair score-level comparison yet.

Exaone 4.0 32B has the larger context window at 128K, compared with 64K for Sarvam 30B.

Category breakdown

Exact category averages are shown below. Not measured means BenchLM does not have enough sourced public coverage for that model and category.

Category scores and score margins for Exaone 4.0 32B and Sarvam 30B
CategoryExaone 4.0 32BΔSarvam 30B
KnowledgeExaone 4.0 32B81.8MarginNo overlapSarvam 30BNot measured

Operational comparison

Runtime and commercial metrics are compared only when both models have a complete sourced value.

MetricExaone 4.0 32BSarvam 30BComparison
Input / output priceUSD per 1M tokensExaone 4.0 32BNot availableSarvam 30B$0 input / $0 outputA complete price comparison is not available.
Generation speedtokens per secondExaone 4.0 32BNot availableSarvam 30BNot availableA complete speed comparison is not available.
First-answer latencyseconds to first tokenExaone 4.0 32BNot availableSarvam 30BNot availableA complete latency comparison is not available.
Context windowmaximum listed tokensExaone 4.0 32B128KSarvam 30B64KExaone 4.0 32B lists the larger context window.

Benchmark Deep Dive

Agentic
BenchmarkExaone 4.0 32BSarvam 30BResult
τ²-bench resultsSource 4.1%34.5%Sarvam 30B leads
Coding
BenchmarkExaone 4.0 32BSarvam 30BResult
AA-SciCodeSource 25.2%19.2%Exaone 4.0 32B leads
Reasoning
BenchmarkExaone 4.0 32BSarvam 30BResult
AA-LCRSource 8.0%0.0%Exaone 4.0 32B leads
CritPtSource 0.0%0.3%Sarvam 30B leads
Knowledge
BenchmarkExaone 4.0 32BSarvam 30BResult
MMLU-ProSource 81.8%Not comparable
Artificial Analysis Intelligence IndexSource 6.0%6.6%Sarvam 30B leads
AA-GPQA DiamondSource 62.8%63.3%Sarvam 30B leads
AA-HLESource 4.9%7.0%Sarvam 30B leads
AA-Omniscience IndexSource -62.3%-72.0%Exaone 4.0 32B leads
AA-Omniscience AccuracySource 10.4%12.7%Sarvam 30B leads
AA-Omniscience Hallucination RateSource 81.0%97.0%Exaone 4.0 32B leads
Math
BenchmarkExaone 4.0 32BSarvam 30BResult
AIME 2025Source 85.3%Not comparable
Inst. Following
BenchmarkExaone 4.0 32BSarvam 30BResult
AA-IFBenchSource 33.5%26.5%Exaone 4.0 32B leads
Frequently Asked Questions (3)

Can I compare Exaone 4.0 32B and Sarvam 30B on BenchLM yet?

Not fully yet. BenchLM is tracking both models, but the sourced benchmark breakdown for this comparison is still coming soon.

Why does this comparison show “coming soon”?

BenchLM only shows category winners and benchmark-level calls when we have sourced results that can be compared fairly. For these models, the public benchmark coverage is not complete enough yet.

What data is available for Exaone 4.0 32B and Sarvam 30B today?

Sarvam 30B: $0.00 input / $0.00 output per 1M tokens Both model pages still include creator, context window, reasoning mode, and other metadata while benchmark coverage fills in.

Related Comparisons

Last updated: July 23, 2026

Choose a model with this week’s evidence

Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.

One email each week. Unsubscribe anytime.