Claude Sonnet 4 vs Qwen3 14B
Wins 6 of 6 areas
Coding · Agents · Reasoning · Facts · Long documents · Following instructions
Wins 0 of 6 areas
—
Claude Sonnet 4 is the stronger all-rounder.Qwen3 14B is cheaper.
Scores updated · 27 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.
- CodingWriting and fixing software30Claude Sonnet 43 of 3 tests
- ReasoningHard problems that need careful thinking20Claude Sonnet 42 of 3 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own20Claude Sonnet 42 of 2 tests
- FactsGetting facts right instead of making them up21Claude Sonnet 42 of 3 tests
- Long documentsFinding answers in very long texts10Claude Sonnet 41 of 1 test
- Following instructionsDoing exactly what it is asked10Claude Sonnet 41 of 1 test
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.
Qwen3 14B costs 90% less for the same work.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where Claude Sonnet 4 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+46.7points ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+31.4points ahead
- Hard command-line tasks in a real terminalTerminal-Bench Hard+27.3points ahead
Where Qwen3 14B pulls ahead
No clear win on a test scored out of 100.
Every test, side by side
All 27 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Sonnet 4
- Terminal-Bench 2.1Claude Sonnet 4 by 31.436.34.9+31.4
- Terminal-Bench HardClaude Sonnet 4 by 27.331.13.8+27.3
- SciCodeClaude Sonnet 4 by 9.34030.7+9.3
AgentsClaude Sonnet 4
- τ-Bench V3 · BankingClaude Sonnet 4 by 11.116.75.6+11.1
- GDPValClaude Sonnet 4 by 9.79.70+9.7
ReasoningClaude Sonnet 4
- GPQA DiamondClaude Sonnet 4 by 17.377.760.4+17.3
- Humanity's Last ExamClaude Sonnet 4 by 6.210.74.5+6.2
- CritPttie0.30tie
FactsClaude Sonnet 4
- AA-Omniscience · Non-hallucinationClaude Sonnet 4 by 46.770.924.2+46.7
- AA-Omniscience · AccuracyClaude Sonnet 4 by 7.422.715.3+7.4
- Vectara HHEM hallucination ratelower is betterQwen3 14B by 4.910.35.4+4.9
Following instructionsClaude Sonnet 4
- IFBenchClaude Sonnet 4 by 14.254.740.5+14.2
Other results14 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.
- AA-OmniscienceClaude Sonnet 4 by 49.10.2-49+49.1
- AIME 2024Qwen3 14B by 40.343.483.7+40.3
- τ²-Bench Telecom (AA run)Claude Sonnet 4 by 30.164.634.5+30.1
- Artificial Analysis Coding IndexClaude Sonnet 4 by 23.837.613.8+23.8
- ZebraLogicQwen3 14B by 14.873.788.5+14.8
- AA IntelligenceClaude Sonnet 4 by 10.718.98.2+10.7
- LiveCodeBenchClaude Sonnet 4 by 9.268.559.3+9.2
- MMLU-ReduxClaude Sonnet 4 by 593.688.6+5
- vectara_factual_consistencyQwen3 14B by 4.989.794.6+4.9
- vectara_avg_summary_lengthClaude Sonnet 4 by 34.7 rating points145.8111.1+34.7 rating
- MATH-500 (EM)Qwen3 14B by 2.89496.8+2.8
- vectara_answer_rateQwen3 14B by 1.398.699.9+1.3
- AutoLogitie89.889.2tie
- AIME 2025tie70.570.4tie
Questions people ask
Which is better, Claude Sonnet 4 or Qwen3 14B?
Claude Sonnet 4 wins all six areas where both have results: coding, agents, reasoning, facts, long documents and following instructions. Qwen3 14B wins none, but costs 90% less.
Which is better for coding?
Claude Sonnet 4. It wins 3 of the 3 coding tests both models report; Qwen3 14B wins none.
Which is cheaper?
Claude Sonnet 4 costs $3.00 per million input tokens and $15.00 per million output tokens; Qwen3 14B costs $0.35 and $1.40. That makes Qwen3 14B about 90% cheaper for the same work.
How do you compare the two?
We use the 27 benchmark tests both models have published scores on. The verdict counts the 13 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 14 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.