GPT-4o vs Qwen3 Next 80B A3B
Wins 1 of 6 areas
Facts
Wins 4 of 6 areas
Coding · Reasoning · Long documents · Following instructions
Qwen3 Next 80B A3B is the stronger all-rounder.GPT-4o is better at facts.
Scores updated · 36 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 Next 80B A3B3 of 3 tests
- ReasoningHard problems that need careful thinking02Qwen3 Next 80B A3B2 of 3 tests · 1 tie
- Long documentsFinding answers in very long texts01Qwen3 Next 80B A3B1 of 1 test
- Following instructionsDoing exactly what it is asked01Qwen3 Next 80B A3B1 of 1 test
- FactsGetting facts right instead of making them up20GPT-4o2 of 3 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own00Even0 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.
Qwen3 Next 80B A3B costs 93% 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 Qwen3 Next 80B A3B pulls ahead
- Recent programming contest problemsLiveCodeBench v6+37.8points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+24.7points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+23.3points ahead
Where GPT-4o pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+49.1points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+4.7points ahead
Every test, side by side
All 36 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingQwen3 Next 80B A3B
- LiveCodeBench v6Qwen3 Next 80B A3B by 37.830.968.7+37.8
- SciCodeQwen3 Next 80B A3B by 5.533.338.8+5.5
- Terminal-Bench HardQwen3 Next 80B A3B by 1.58.39.8+1.5
ReasoningQwen3 Next 80B A3B
- GPQA DiamondQwen3 Next 80B A3B by 23.352.675.9+23.3
- Humanity's Last ExamQwen3 Next 80B A3B by 10.81.812.6+10.8
- CritPttie00tie
FactsGPT-4o
- AA-Omniscience · Non-hallucinationGPT-4o by 49.162.113+49.1
- AA-Omniscience · AccuracyGPT-4o by 4.723.719+4.7
- Vectara HHEM hallucination ratelower is bettertie9.69.3tie
Long documentsQwen3 Next 80B A3B
- AA-LCRQwen3 Next 80B A3B by 7.75663.7+7.7
Following instructionsQwen3 Next 80B A3B
- IFBenchQwen3 Next 80B A3B by 24.73660.7+24.7
Other results24 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 Next 80B A3B by 76.111.787.8+76.1
- AA-OmniscienceGPT-4o by 40.9-10.5-51.4+40.9
- AIR-Bench 2024Qwen3 Next 80B A3B by 24.362.486.7+24.3
- SuperGPQAQwen3 Next 80B A3B by 18.442.460.8+18.4
- MMLU-ProXQwen3 Next 80B A3B by 17.661.178.7+17.6
- vectara_avg_summary_lengthGPT-4o by 15.786.670.9+15.7
- Tau2 airlineQwen3 Next 80B A3B by 1545.560.5+15
- τ²-Bench Telecom (AA run)Qwen3 Next 80B A3B by 12.628.941.5+12.6
- TAU-bench (retail)Qwen3 Next 80B A3B by 9.360.369.6+9.3
- MMLU-ProQwen3 Next 80B A3B by 874.782.7+8
- IFEvalQwen3 Next 80B A3B by 7.98188.9+7.9
- Artificial Analysis Coding IndexGPT-4o by 6.824.217.4+6.8
- AA Agentic IndexGPT-4o by 6.38.42.1+6.3
- TAU-bench (airline)Qwen3 Next 80B A3B by 6.242.849+6.2
- MMLU-ReduxQwen3 Next 80B A3B by 4.58892.5+4.5
- τ²-Bench (Retail)Qwen3 Next 80B A3B by 4.463.467.8+4.4
- AA IntelligenceQwen3 Next 80B A3B by 3.97.311.2+3.9
- bbqGPT-4o by 3.495.191.7+3.4
- HarmBenchQwen3 Next 80B A3B by 1.782.984.6+1.7
- simple_safety_testsQwen3 Next 80B A3B by 198.599.5+1
- XSTestQwen3 Next 80B A3B by 197.398.3+1
- anthropic_red_teamtie99.199.9tie
- vectara_answer_ratetie93.894.4tie
- vectara_factual_consistencytie90.490.7tie
Questions people ask
Which is better, GPT-4o or Qwen3 Next 80B A3B?
Qwen3 Next 80B A3B wins four of the six areas where both have results: coding, reasoning, long documents and following instructions. GPT-4o wins facts. They are level on agents.
Which is better for coding?
Qwen3 Next 80B A3B. 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 Next 80B A3B costs $0.15 and $1.20. That makes Qwen3 Next 80B A3B about 93% cheaper for the same work.
How do you compare the two?
We use the 36 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 24 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.