Model comparison
Gemma 4 E4B vs Sarvam 105B
Head-to-head evidence from 11 shared benchmark results across 5 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.
Public leaderboard positions: Gemma 4 E4B #155 (Estimated); Sarvam 105B #157 (Estimated). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.
Evidence parity. Gemma 4 E4B and Sarvam 105B share 11 comparable benchmark results. 0 of 8 categories are comparable. 3 results are unique to Gemma 4 E4B; 0 to Sarvam 105B.
Updated July 23, 2026- Shared results
- 11
- Gemma 4 E4B only
- 3
- Sarvam 105B only
- 0
- Comparable categories
- 0 / 8
Benchmark data for Gemma 4 E4B and Sarvam 105B 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.
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 | Gemma 4 E4B | Δ | Sarvam 105B |
|---|---|---|---|
| Knowledge | Gemma 4 E4B67.4 | MarginNo overlap | Sarvam 105BNot measured |
Operational comparison
Runtime and commercial metrics are compared only when both models have a complete sourced value.
| Metric | Gemma 4 E4B | Sarvam 105B | Comparison |
|---|---|---|---|
| Input / output priceUSD per 1M tokens | Gemma 4 E4B$0 input / $0 output | Sarvam 105B$0 input / $0 output | Listed prices are equal. |
| Generation speedtokens per second | Gemma 4 E4BNot available | Sarvam 105BNot available | A complete speed comparison is not available. |
| First-answer latencyseconds to first token | Gemma 4 E4BNot available | Sarvam 105BNot available | A complete latency comparison is not available. |
| Context windowmaximum listed tokens | Gemma 4 E4B128K | Sarvam 105B128K | Listed context windows are equal. |
Benchmark Deep Dive
Agentic1 benchmarks
| Benchmark | Gemma 4 E4B | Sarvam 105B | Result |
|---|---|---|---|
| τ²-bench resultsSource | 20.8% | 46.8% | Sarvam 105B leads |
Coding1 benchmarks
| Benchmark | Gemma 4 E4B | Sarvam 105B | Result |
|---|---|---|---|
| AA-SciCodeSource | 24.4% | 26.4% | Sarvam 105B leads |
Reasoning2 benchmarks
Knowledge8 benchmarks
| Benchmark | Gemma 4 E4B | Sarvam 105B | Result |
|---|---|---|---|
| GPQASource | 58.6% | — | Not comparable |
| MMLU-ProSource | 69.4% | — | Not comparable |
| Artificial Analysis Intelligence IndexSource | 12.5% | 11.9% | Gemma 4 E4B leads |
| AA-GPQA DiamondSource | 57.6% | 73.8% | Sarvam 105B leads |
| AA-HLESource | 3.7% | 10.1% | Sarvam 105B leads |
| AA-Omniscience IndexSource | -20.0% | -59.5% | Gemma 4 E4B leads |
| AA-Omniscience AccuracySource | 8.6% | 17.6% | Sarvam 105B leads |
| AA-Omniscience Hallucination RateSource | 31.3% | 93.5% | Gemma 4 E4B leads |
Multimodal1 benchmarks
| Benchmark | Gemma 4 E4B | Sarvam 105B | Result |
|---|---|---|---|
| AA-MMMU-ProSource | 51.4% | — | Not comparable |
Inst. Following1 benchmarks
| Benchmark | Gemma 4 E4B | Sarvam 105B | Result |
|---|---|---|---|
| AA-IFBenchSource | 44.2% | 34.4% | Gemma 4 E4B leads |
Frequently Asked Questions (3)
Can I compare Gemma 4 E4B and Sarvam 105B 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 Gemma 4 E4B and Sarvam 105B today?
Gemma 4 E4B: $0.00 input / $0.00 output per 1M tokens Sarvam 105B: $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
Explore More
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.