GPT-5.6 Luna vs Qwen3.6 35B A3B
Wins 6 of 7 areas
Coding · Agents · Reasoning · Images and charts · Math · Long documents
Wins 0 of 7 areas
—
GPT-5.6 Luna is the stronger all-rounder.
Scores updated · 36 tests both models report · How we compare
Where each one wins
Tests won in each of the seven 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 software40GPT-5.6 Luna4 of 4 tests
- AgentsCarrying out multi-step tasks on its own40GPT-5.6 Luna4 of 4 tests
- ReasoningHard problems that need careful thinking30GPT-5.6 Luna3 of 3 tests
- MathCompetition and research-level math30GPT-5.6 Luna3 of 3 tests
- Images and chartsUnderstanding pictures, charts and video21GPT-5.6 Luna2 of 3 tests
- Long documentsFinding answers in very long texts10GPT-5.6 Luna1 of 1 test
- FactsGetting facts right instead of making them up11Even1 each
Long documents rests on a single test.
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.
GPT-5.6 Luna costs 47% 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 GPT-5.6 Luna pulls ahead
- Unpublished advanced math problemsFrontierMath Tiers 1-3 (v2)+61.7points ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+36points ahead
- Finds hard-to-locate facts by browsing the webBrowseComp+34.6points ahead
Where Qwen3.6 35B A3B pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+42.1points ahead
- Math problems shown in pictures and chartsMathVista+9.5points 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.
CodingGPT-5.6 Luna
- Terminal-Bench 2.1GPT-5.6 Luna by 3680.944.9+36
- SWE-bench VerifiedGPT-5.6 Luna by 19.69373.4+19.6
- SciCodeGPT-5.6 Luna by 1753.636.6+17
- SWE-bench ProGPT-5.6 Luna by 13.262.749.5+13.2
AgentsGPT-5.6 Luna
- BrowseCompGPT-5.6 Luna by 34.683.348.7+34.6
- GDPValGPT-5.6 Luna by 28.44819.6+28.4
- τ-Bench V3 · BankingGPT-5.6 Luna by 21.831.19.3+21.8
- Terminal-Bench 4.0GPT-5.6 Luna by 11.611.60+11.6
ReasoningGPT-5.6 Luna
- CritPtGPT-5.6 Luna by 20.320.60.3+20.3
- Humanity's Last ExamGPT-5.6 Luna by 17.339.522.2+17.3
- GPQA DiamondGPT-5.6 Luna by 791.184.1+7
FactsEven
- AA-Omniscience · Non-hallucinationQwen3.6 35B A3B by 42.17.449.5+42.1
- AA-Omniscience · AccuracyGPT-5.6 Luna by 23.942.718.8+23.9
Images and chartsGPT-5.6 Luna
- MathVistaQwen3.6 35B A3B by 9.576.986.4+9.5
- CharXiv (RQ)GPT-5.6 Luna by 4.782.778+4.7
- MMMU-ProGPT-5.6 Luna by 3.678.675+3.6
MathGPT-5.6 Luna
- FrontierMath Tiers 1-3 (v2)GPT-5.6 Luna by 61.782.120.4+61.7
- HMMT Feb. 2026GPT-5.6 Luna by 14.998.583.6+14.9
- AIME 2026GPT-5.6 Luna by 4.997.692.7+4.9
Other results16 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.
- GDPval-AA v2 EloGPT-5.6 Luna by 1038 rating points1584546+1038 rating
- GDPval-AA v2GPT-5.6 Luna by 515 rating points15301015+515 rating
- Artificial Analysis Coding IndexGPT-5.6 Luna by 29.571.441.9+29.5
- AA Agentic IndexGPT-5.6 Luna by 27.742.715+27.7
- ToolathlonGPT-5.6 Luna by 26.553.426.9+26.5
- AA IntelligenceGPT-5.6 Luna by 19.137.318.2+19.1
- τ³-Bench BankingGPT-5.6 Luna by 16.724.37.6+16.7
- RealWorldQAQwen3.6 35B A3B by 12.972.485.3+12.9
- LVBenchQwen3.6 35B A3B by 12.558.971.4+12.5
- AA-OmniscienceGPT-5.6 Luna by 11.9-10.3-22.2+11.9
- RefSpatial-BenchQwen3.6 35B A3B by 11.852.564.3+11.8
- HallusionBenchQwen3.6 35B A3B by 10.85969.8+10.8
- EmbSpatial-BenchQwen3.6 35B A3B by 8.875.584.3+8.8
- MV-BenchQwen3.6 35B A3B by 6.468.274.6+6.4
- SimpleVQAQwen3.6 35B A3B by 5.753.258.9+5.7
- VideoMMMUQwen3.6 35B A3B by 2.681.183.7+2.6
Questions people ask
Which is better, GPT-5.6 Luna or Qwen3.6 35B A3B?
GPT-5.6 Luna wins six of the seven areas where both have results: coding, agents, reasoning, images and charts, math and long documents. Qwen3.6 35B A3B wins none. They are level on facts.
Which is better for coding?
GPT-5.6 Luna. It wins 4 of the 4 coding tests both models report; Qwen3.6 35B A3B wins none.
Which is cheaper?
GPT-5.6 Luna costs $0.20 per million input tokens and $1.20 per million output tokens; Qwen3.6 35B A3B costs $0.38 and $2.25. That makes GPT-5.6 Luna about 47% 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 20 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 16 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.