Qwen3 235B A22B vs Sarvam 105B
Wins 2 of 6 areas
Facts · Following instructions
Wins 1 of 6 areas
Reasoning
Qwen3 235B A22B wins more areas, narrowly.Sarvam 105B is cheaper and better at reasoning.
Scores updated · 22 tests both models report · How we compare
Where each one wins
Tests won in each of the six areas where both have results. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
- FactsGetting facts right instead of making them up10Qwen3 235B A22B1 of 2 tests · 1 tie
- Following instructionsDoing exactly what it is asked10Qwen3 235B A22B1 of 1 test
- ReasoningHard problems that need careful thinking01Sarvam 105B1 of 3 tests · 2 ties
- CodingWriting and fixing software22Even2 each
- AgentsCarrying out multi-step tasks on its own00Even0 each · 1 tie
- Long documentsFinding answers in very long texts00Even0 each · 1 tie
Agents, long documents and following instructions rest on a single test each.
What it costs
Prices per million tokens, roughly 750,000 words. The bars show the cost of a million tokens read plus a million written.
Sarvam 105B costs 94% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Qwen3 235B A22B pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+15.8points ahead
- Code for real scientific research problemsSciCode+13.5points ahead
- Hard command-line tasks in a real terminalTerminal-Bench Hard+4.6points ahead
Where Sarvam 105B pulls ahead
- Recent programming contest problemsLiveCodeBench v6+34.7points ahead
- Fixes real GitHub issues in Python projectsSWE-bench Verified+10.6points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+3.8points ahead
Every test, side by side
All 22 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingEven
- LiveCodeBench v6Sarvam 105B by 34.73771.7+34.7
- SciCodeQwen3 235B A22B by 13.539.926.4+13.5
- SWE-bench VerifiedSarvam 105B by 10.634.445+10.6
- Terminal-Bench HardQwen3 235B A22B by 4.66.11.5+4.6
ReasoningSarvam 105B
- GPQA DiamondSarvam 105B by 3.87073.8+3.8
- CritPttie00tie
- Humanity's Last Examtie1111tie
FactsQwen3 235B A22B
- AA-Omniscience · Non-hallucinationQwen3 235B A22B by 15.822.46.6+15.8
- AA-Omniscience · Accuracytie18.517.6tie
Following instructionsQwen3 235B A22B
- IFBenchQwen3 235B A22B by 4.338.734.4+4.3
Other results10 tests, not counted
Tests outside the eight areas. They are not counted above: several are summary scores built from other tests, or the same test under another name.
- HMMT 2025Sarvam 105B by 73.911.985.8+73.9
- τ²-Bench Telecom (AA run)Sarvam 105B by 22.82446.8+22.8
- AIME 2025Sarvam 105B by 15.281.596.7+15.2
- AA-OmniscienceQwen3 235B A22B by 14.7-44.7-59.4+14.7
- MMLU-ProSarvam 105B by 13.568.281.7+13.5
- Artificial Analysis Coding IndexQwen3 235B A22B by 7.617.49.8+7.6
- MATH-500 (EM)Sarvam 105B by 7.491.298.6+7.4
- AA Agentic IndexSarvam 105B by 6.318.424.7+6.3
- IFEvalSarvam 105B by 1.683.284.8+1.6
- AA Intelligencetie9.58.8tie
Questions people ask
Which is better, Qwen3 235B A22B or Sarvam 105B?
Qwen3 235B A22B wins two of the six areas where both have results: facts and following instructions. Sarvam 105B wins reasoning, and costs 94% less. They are level on coding, agents and long documents.
Which is better for coding?
Neither. They win 2 coding tests each of the 4 both models report.
Which is cheaper?
Qwen3 235B A22B costs $0.70 per million input tokens and $2.80 per million output tokens; Sarvam 105B costs $0.04 and $0.17. That makes Sarvam 105B about 94% cheaper for the same work.
How do you compare the two?
We use the 22 benchmark tests both models have published scores on. The verdict counts the 12 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 10 are listed but not counted, because several are summary scores or repeat a test. Each score is the one shown on the model's own page.