Qwen3.5 122B A10B vs Seed 1.8
Wins 1 of 4 areas
Images and charts
Wins 2 of 4 areas
Agents · Following instructions
Seed 1.8 wins more areas, narrowly.Qwen3.5 122B A10B is better at images and charts.
Scores updated · 27 tests both models report · How we compare
Where each one wins
Tests won in each of the four 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.
Following instructions 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.
Seed 1.8 costs 38% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Seed 1.8 pulls ahead
- Keeps track of context across a multi-turn chatMulti-Challenge+5.2points ahead
- Completes tasks by operating a computer desktopOSWorld-Verified+3.9points ahead
- Finds hard-to-locate facts by browsing the webBrowseComp+3.8points ahead
Where Qwen3.5 122B A10B pulls ahead
- Reasoning about charts from research papersCharXiv (RQ)+5.8points ahead
- Questions that truly need the image to answerMMStar+3points ahead
- Harder college exam questions with imagesMMMU-Pro+1.8points ahead
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.
AgentsSeed 1.8
- OSWorld-VerifiedSeed 1.8 by 3.95861.9+3.9
- BrowseCompSeed 1.8 by 3.863.867.6+3.8
Images and chartsQwen3.5 122B A10B
- CharXiv (RQ)Qwen3.5 122B A10B by 5.877.271.4+5.8
- MMStarQwen3.5 122B A10B by 382.979.9+3
- MMMU-ProQwen3.5 122B A10B by 1.87573.2+1.8
- MMMUtie83.983.4tie
- MathVistatie87.487.7tie
Following instructionsSeed 1.8
- Multi-ChallengeSeed 1.8 by 5.261.566.7+5.2
Other results17 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.
- DynaMathQwen3.5 122B A10B by 24.485.961.5+24.4
- BFCLv4Qwen3.5 122B A10B by 1572.257.2+15
- RefSpatial-BenchQwen3.5 122B A10B by 1369.356.3+13
- BrowseComp-ZHSeed 1.8 by 11.469.981.3+11.4
- MATH-VisionQwen3.5 122B A10B by 4.986.281.3+4.9
- AI2DQwen3.5 122B A10B by 4.293.389.1+4.2
- Terminal-Bench 2.0Qwen3.5 122B A10B by 4.249.445.2+4.2
- HallusionBenchQwen3.5 122B A10B by 3.767.663.9+3.7
- SimpleVQASeed 1.8 by 3.761.765.4+3.7
- WideSearchSeed 1.8 by 3.360.563.8+3.3
- ERQAQwen3.5 122B A10B by 3.26258.8+3.2
- SuperGPQAQwen3.5 122B A10B by 2.367.164.8+2.3
- ZeroBenchSeed 1.8 by 2911+2
- MMLU-ProQwen3.5 122B A10B by 1.886.784.9+1.8
- MMVUQwen3.5 122B A10B by 1.674.773.1+1.6
- LVBenchQwen3.5 122B A10B by 1.474.473+1.4
- VideoMMMUtie8282.7tie
Questions people ask
Which is better, Qwen3.5 122B A10B or Seed 1.8?
Seed 1.8 wins two of the four areas where both have results: agents and following instructions. Qwen3.5 122B A10B wins images and charts. They are level on coding.
Which is better for coding?
Neither. All 2 coding tests both models report are ties.
Which is cheaper?
Qwen3.5 122B A10B costs $0.40 per million input tokens and $3.20 per million output tokens; Seed 1.8 costs $0.25 and $2.00. That makes Seed 1.8 about 38% 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 10 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 17 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.