Qwen3 235B A22B Instruct 2507 vs Qwen3 Next 80B A3B Instruct
Wins 5 of 6 areas
Coding · Agents · Reasoning · Facts · Following instructions
Wins 1 of 6 areas
Long documents
Qwen3 235B A22B Instruct 2507 is the stronger all-rounder.Qwen3 Next 80B A3B Instruct is better at long documents.
Scores updated · 30 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.
- ReasoningHard problems that need careful thinking20Qwen3 235B A22B Instruct 25072 of 3 tests · 1 tie
- FactsGetting facts right instead of making them up20Qwen3 235B A22B Instruct 25072 of 2 tests
- CodingWriting and fixing software21Qwen3 235B A22B Instruct 25072 of 3 tests
- AgentsCarrying out multi-step tasks on its own10Qwen3 235B A22B Instruct 25071 of 1 test
- Following instructionsDoing exactly what it is asked10Qwen3 235B A22B Instruct 25071 of 1 test
- Long documentsFinding answers in very long texts01Qwen3 Next 80B A3B Instruct1 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 15% 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 235B A22B Instruct 2507 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+15.5points ahead
- Real work tasks from 44 professionsGDPVal+7.7points ahead
- Hard command-line tasks in a real terminalTerminal-Bench Hard+7.6points ahead
Where Qwen3 Next 80B A3B Instruct pulls ahead
- Reasons across sets of long documentsAA-LCR+18.8points ahead
- Recent programming contest problemsLiveCodeBench v6+4.8points ahead
Every test, side by side
All 30 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingQwen3 235B A22B Instruct 2507
- Terminal-Bench HardQwen3 235B A22B Instruct 2507 by 7.615.27.6+7.6
- SciCodeQwen3 235B A22B Instruct 2507 by 5.33630.7+5.3
- LiveCodeBench v6Qwen3 Next 80B A3B Instruct by 4.851.856.6+4.8
AgentsQwen3 235B A22B Instruct 2507
- GDPValQwen3 235B A22B Instruct 2507 by 7.714.16.4+7.7
ReasoningQwen3 235B A22B Instruct 2507
- Humanity's Last ExamQwen3 235B A22B Instruct 2507 by 3.511.17.6+3.5
- GPQA DiamondQwen3 235B A22B Instruct 2507 by 1.575.373.8+1.5
- CritPttie00tie
FactsQwen3 235B A22B Instruct 2507
- AA-Omniscience · Non-hallucinationQwen3 235B A22B Instruct 2507 by 15.522.87.3+15.5
- AA-Omniscience · AccuracyQwen3 235B A22B Instruct 2507 by 1.418.717.3+1.4
Long documentsQwen3 Next 80B A3B Instruct
- AA-LCRQwen3 Next 80B A3B Instruct by 18.833.952.7+18.8
Following instructionsQwen3 235B A22B Instruct 2507
- IFBenchQwen3 235B A22B Instruct 2507 by 6.446.139.7+6.4
Other results19 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-OmniscienceQwen3 235B A22B Instruct 2507 by 15.5-44-59.5+15.5
- τ²-Bench (Retail)Qwen3 235B A22B Instruct 2507 by 1471.357.3+14
- τ²-Bench Telecom (AA run)Qwen3 235B A22B Instruct 2507 by 11.733.321.6+11.7
- AA Agentic IndexQwen3 235B A22B Instruct 2507 by 8.622.814.2+8.6
- Aider-PolyglotQwen3 235B A22B Instruct 2507 by 7.557.349.8+7.5
- Artificial Analysis Coding IndexQwen3 235B A22B Instruct 2507 by 6.822.115.3+6.8
- SuperGPQAQwen3 235B A22B Instruct 2507 by 3.862.658.8+3.8
- Arena-Hard-V2Qwen3 Next 80B A3B Instruct by 3.579.282.7+3.5
- MMLU-ProXQwen3 235B A22B Instruct 2507 by 2.779.476.7+2.7
- AA IntelligenceQwen3 235B A22B Instruct 2507 by 2.4129.6+2.4
- MMLU-ProQwen3 235B A22B Instruct 2507 by 2.48380.6+2.4
- MMLU-ReduxQwen3 235B A22B Instruct 2507 by 2.293.190.9+2.2
- Tau2 airlineQwen3 Next 80B A3B Instruct by 1.54445.5+1.5
- HMMT 2025Qwen3 235B A22B Instruct 2507 by 1.355.454.1+1.3
- IFEvalQwen3 235B A22B Instruct 2507 by 1.188.787.6+1.1
- AIME 2025tie70.369.5tie
- BFCL v3tie70.970.3tie
- Includetie79.578.9tie
- MultiPL-Etie87.987.8tie
Questions people ask
Which is better, Qwen3 235B A22B Instruct 2507 or Qwen3 Next 80B A3B Instruct?
Qwen3 235B A22B Instruct 2507 wins five of the six areas where both have results: coding, agents, reasoning, facts and following instructions. Qwen3 Next 80B A3B Instruct wins long documents.
Which is better for coding?
Qwen3 235B A22B Instruct 2507. It wins 2 of the 3 coding tests both models report; Qwen3 Next 80B A3B Instruct wins 1.
Which is cheaper?
Qwen3 235B A22B Instruct 2507 costs $0.23 per million input tokens and $0.92 per million output tokens; Qwen3 Next 80B A3B Instruct costs $0.15 and $1.20. That makes Qwen3 235B A22B Instruct 2507 about 15% cheaper for the same work.
How do you compare the two?
We use the 30 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 19 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.