GPT-4o vs Qwen3 235B A22B Instruct 2507
Wins 2 of 6 areas
Facts · Long documents
Wins 4 of 6 areas
Coding · Agents · Reasoning · Following instructions
Qwen3 235B A22B Instruct 2507 is the stronger all-rounder.GPT-4o is better at facts.
Scores updated · 33 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 software03Qwen3 235B A22B Instruct 25073 of 3 tests
- ReasoningHard problems that need careful thinking02Qwen3 235B A22B Instruct 25072 of 3 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own01Qwen3 235B A22B Instruct 25071 of 1 test
- Following instructionsDoing exactly what it is asked01Qwen3 235B A22B Instruct 25071 of 1 test
- FactsGetting facts right instead of making them up20GPT-4o2 of 2 tests
- Long documentsFinding answers in very long texts10GPT-4o1 of 1 test
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.
Qwen3 235B A22B Instruct 2507 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 Instruct 2507 pulls ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+22.7points ahead
- Recent programming contest problemsLiveCodeBench v6+20.9points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+10.1points ahead
Where GPT-4o pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+39.3points ahead
- Reasons across sets of long documentsAA-LCR+22.1points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+5points ahead
Every test, side by side
All 33 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingQwen3 235B A22B Instruct 2507
- LiveCodeBench v6Qwen3 235B A22B Instruct 2507 by 20.930.951.8+20.9
- Terminal-Bench HardQwen3 235B A22B Instruct 2507 by 6.98.315.2+6.9
- SciCodeQwen3 235B A22B Instruct 2507 by 2.733.336+2.7
AgentsQwen3 235B A22B Instruct 2507
- GDPValQwen3 235B A22B Instruct 2507 by 14.1014.1+14.1
ReasoningQwen3 235B A22B Instruct 2507
- GPQA DiamondQwen3 235B A22B Instruct 2507 by 22.752.675.3+22.7
- Humanity's Last ExamQwen3 235B A22B Instruct 2507 by 9.31.811.1+9.3
- CritPttie00tie
FactsGPT-4o
- AA-Omniscience · Non-hallucinationGPT-4o by 39.362.122.8+39.3
- AA-Omniscience · AccuracyGPT-4o by 523.718.7+5
Following instructionsQwen3 235B A22B Instruct 2507
- IFBenchQwen3 235B A22B Instruct 2507 by 10.13646.1+10.1
Other results22 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.
- AIME 2025Qwen3 235B A22B Instruct 2507 by 58.611.770.3+58.6
- AA-OmniscienceGPT-4o by 33.5-10.5-44+33.5
- Aider-PolyglotQwen3 235B A22B Instruct 2507 by 26.630.757.3+26.6
- SuperGPQAQwen3 235B A22B Instruct 2507 by 20.242.462.6+20.2
- MMLU-ProXQwen3 235B A22B Instruct 2507 by 18.361.179.4+18.3
- AIR-Bench 2024Qwen3 235B A22B Instruct 2507 by 16.662.479+16.6
- SimpleQAQwen3 235B A22B Instruct 2507 by 16.138.254.3+16.1
- AA Agentic IndexQwen3 235B A22B Instruct 2507 by 14.48.422.8+14.4
- MMLU-ProQwen3 235B A22B Instruct 2507 by 8.374.783+8.3
- τ²-Bench (Retail)Qwen3 235B A22B Instruct 2507 by 7.963.471.3+7.9
- IFEvalQwen3 235B A22B Instruct 2507 by 7.78188.7+7.7
- ARC-AGI-1Qwen3 235B A22B Instruct 2507 by 6.54.511+6.5
- MMLU-ReduxQwen3 235B A22B Instruct 2507 by 5.18893.1+5.1
- AA IntelligenceQwen3 235B A22B Instruct 2507 by 4.77.312+4.7
- τ²-Bench Telecom (AA run)Qwen3 235B A22B Instruct 2507 by 4.428.933.3+4.4
- HarmBenchGPT-4o by 3.382.979.6+3.3
- Artificial Analysis Coding IndexGPT-4o by 2.124.222.1+2.1
- simple_safety_testsQwen3 235B A22B Instruct 2507 by 1.598.5100+1.5
- Tau2 airlineGPT-4o by 1.545.544+1.5
- XSTestQwen3 235B A22B Instruct 2507 by 1.397.398.6+1.3
- bbqQwen3 235B A22B Instruct 2507 by 1.195.196.2+1.1
- anthropic_red_teamtie99.199.9tie
Questions people ask
Which is better, GPT-4o or Qwen3 235B A22B Instruct 2507?
Qwen3 235B A22B Instruct 2507 wins four of the six areas where both have results: coding, agents, reasoning and following instructions. GPT-4o wins facts and long documents.
Which is better for coding?
Qwen3 235B A22B Instruct 2507. It wins 3 of the 3 coding tests both models report; GPT-4o wins none.
Which is cheaper?
GPT-4o costs $5.00 per million input tokens and $15.00 per million output tokens; Qwen3 235B A22B Instruct 2507 costs $0.23 and $0.92. That makes Qwen3 235B A22B Instruct 2507 about 94% cheaper for the same work.
How do you compare the two?
We use the 33 benchmark tests both models have published scores on. The verdict counts the 11 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 22 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.