Side-by-side benchmark comparison across agentic, coding, multimodal, knowledge, reasoning, and math workflows.
GPT-5.4 mini
66
Winner · 4/8 categoriesSarvam 105B
60
1/8 categoriesGPT-5.4 mini· Sarvam 105B
Pick GPT-5.4 mini 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.
GPT-5.4 mini is clearly ahead on the aggregate, 66 to 60. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
GPT-5.4 mini's sharpest advantage is in agentic, where it averages 65.6 against 49.5. The single biggest benchmark swing on the page is IFEval, 87.4% to 84.8%. Sarvam 105B does hit back in knowledge, so the answer changes if that is the part of the workload you care about most.
GPT-5.4 mini is also the more expensive model on tokens at $0.75 input / $4.50 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. GPT-5.4 mini gives you the larger context window at 400K, compared with 128K for Sarvam 105B.
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.
| Benchmark | GPT-5.4 mini | Sarvam 105B |
|---|---|---|
| AgenticGPT-5.4 mini wins | ||
| Terminal-Bench 2.0 | 60% | — |
| OSWorld-Verified | 72.1% | — |
| MCP Atlas | 57.7% | — |
| Toolathlon | 42.9% | — |
| Tau2-Telecom | 93.4% | — |
| BrowseComp | — | 49.5% |
| CodingGPT-5.4 mini wins | ||
| SWE-bench Pro | 54.4% | — |
| LiveCodeBench v6 | — | 71.7% |
| SWE-bench Verified | — | 45% |
| Multimodal & Grounded | ||
| Coming soon | ||
| Reasoning | ||
| MRCRv2 | 40.7% | — |
| MRCR v2 64K-128K | 47.7% | — |
| MRCR v2 128K-256K | 33.6% | — |
| Graphwalks Parents 128K | 71.5% | — |
| gpqaDiamond | — | 78.7% |
| hle | — | 11.2% |
| KnowledgeSarvam 105B wins | ||
| GPQA | 88% | — |
| HLE | 41.5% | — |
| HLE w/o tools | 28.2% | — |
| MMLU | — | 90.6% |
| MMLU-Pro | — | 81.7% |
| Instruction FollowingGPT-5.4 mini wins | ||
| IFEval | 87.4% | 84.8% |
| Multilingual | ||
| Coming soon | ||
| MathematicsGPT-5.4 mini wins | ||
| MATH-500 | 97.4% | 98.6% |
| AIME 2025 | — | 88.3% |
| HMMT Feb 2025 | — | 85.8% |
| HMMT Nov 2025 | — | 85.8% |
GPT-5.4 mini is ahead overall, 66 to 60. The biggest single separator in this matchup is IFEval, where the scores are 87.4% and 84.8%.
Sarvam 105B has the edge for knowledge tasks in this comparison, averaging 81.7 versus 57.4. GPT-5.4 mini stays close enough that the answer can still flip depending on your workload.
GPT-5.4 mini has the edge for coding in this comparison, averaging 54.4 versus 45. Sarvam 105B stays close enough that the answer can still flip depending on your workload.
GPT-5.4 mini has the edge for math in this comparison, averaging 97.4 versus 92.3. Inside this category, MATH-500 is the benchmark that creates the most daylight between them.
GPT-5.4 mini has the edge for agentic tasks in this comparison, averaging 65.6 versus 49.5. Sarvam 105B stays close enough that the answer can still flip depending on your workload.
GPT-5.4 mini has the edge for instruction following in this comparison, averaging 87.4 versus 84.8. Inside this category, IFEval is the benchmark that creates the most daylight between them.
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.