Qwen3.5 122B A10B — Benchmarks

Benchmark scores for Qwen3.5 122B A10B aggregated from public leaderboards (8 boards), with how it ranks among open models and how it compares with its nearest neighbours. See hardware requirements for what you need to run it.

llmrun Composite Score →

109.4

90% range 102.1–116.8 · from 6 boards

CodingNot enough data yet
AgentsNot enough data yet
MathNot enough data yet
Science96.190% range 74–119
Reasoning112.390% range 104–121

Qwen3.5 122B A10B reasoning benchmarks

Multi-step logic and problem-solving — can the model think through a hard problem rather than recall an answer.

CritPt
CritPt (Critical Points) tests graduate-level physics problem-solving on research-style questions spanning mechanics, electromagnetism, quantum mechanics, and more. It was created by physicists to probe whether a model can reason through multi-step physics rather than pattern-match textbook answers.
0.9%
0.9%
median 0.9%open leader 26.6%
Overall best: Gemini 4 Carbon (high) 36.0%
51st percentile · #41 of 84 open · #174 of 244 all
#41 of 84
#174 of 244
Lab-reported
Epoch AI · Feb 2026
DTBench
DTBench is a set of handcrafted, expert-validated multiple-choice questions on decision theory — Newcomb-like problems and related puzzles — built by the Conceptual Reasoning Index project with Anthropic. It tests whether a model can follow an abstract argument rather than recall a fact.
84.3%
84.3%
median 68.7%open leader 93.9%
Overall best: Claude Opus 5.5 Max 98.9%
75th percentile · #18 of 70 open · #113 of 235 all
#18 of 70
#113 of 235
Lab-reported
Epoch AI · Feb 2026
LMCA
LMCA (Language Model Conceptual Argumentation) scores models on expert-rated arguments across philosophy, decision theory, and AI risk — domains where there is no empirical feedback loop to check an answer against. It is part of the Conceptual Reasoning Index, built with Anthropic by researchers Emery Cooper and Caspar Oesterheld.
32.2%
32.2%
median 25.5%open leader 55.5%
Overall best: Claude Sonnet 5.5 Max 68.3%
67th percentile · #19 of 56 open · #134 of 197 all
#19 of 56
#134 of 197
Lab-reported
Epoch AI · Feb 2026
Mystery Game Puzzles
Mystery Game Puzzles is an Epoch AI evaluation of multi-step deductive reasoning: the model works through a mystery or logic-game scenario to arrive at the correct solution. It probes sustained, structured inference rather than single-step question answering.
17.0%
17.0%
median 18.5%open leader 49.0%
Overall best: GPT 6 Astra Max 84.0%
34th percentile · #14 of 20 open · #80 of 136 all
#14 of 20
#80 of 136
Independent
Epoch AI · Feb 2026

Qwen3.5 122B A10B coding benchmarks

Writing, editing, and fixing real code — measured by whether the resulting program actually runs and passes tests.

SciCode
SciCode asks a model to write real scientific-computing code — implementing multi-step numerical methods drawn from physics, chemistry, biology, and math research — and checks the output against reference solutions. It was built by a team of scientists and ML researchers to test coding ability on genuine research problems, not typical software tasks.
35.6%
35.6%
median 40.1%open leader 60.9%
Overall best: Claude Opus 5.5 Max 66.9%
27th percentile · #53 of 72 open · #193 of 217 all
#53 of 72
#193 of 217
Lab-reported
Epoch AI · Feb 2026

Other results (1)leaderboard-only, not part of the llmrun Score

Leaderboard-only, not part of the llmrun Score

LMArena WebDev
Code Arena (formerly WebDev Arena) is Arena's head-to-head coding leaderboard: two models each build a real web app or UI from the same prompt, and people vote blind on which result is better. The rating shown is a relative Bradley-Terry score, not a percentage — compare it only to other models on this board.
1360.4
1360.4
median 1415.2open leader 1665.0
Overall best: Claude Opus 5.5 (Max) 1809.8
33rd percentile · #37 of 55 open · #112 of 146 all
#37 of 55
#112 of 146
Arena
Arena · Oct 2026

Qwen3.5 122B A10B knowledge benchmarks

Breadth and depth of factual knowledge across many subjects, paired with the reasoning to apply it.

MMLU-Pro
MMLU-Pro is a harder, cleaned-up successor to MMLU with ten answer choices and more reasoning-heavy questions across 14 subjects, measuring broad knowledge and reasoning together.
86.7%
86.7%
median 52.0%open leader 88.0%
Overall best: Gemini-3.1-Pro 91.2%
98th percentile · #4 of 146 open · #21 of 257 all
#4 of 146
#21 of 257
Independent
TIGER-Lab

Qwen3.5 122B A10B instruction following benchmarks

How precisely a model obeys explicit, verifiable instructions — format, length, and content constraints — and, more broadly, how much people prefer its responses head-to-head in blind human voting.

LMArena Text
Arena (arena.ai, formerly LMArena / Chatbot Arena) ranks models by millions of blind, head-to-head human votes on real prompts. The Arena Score shown here is a Bradley-Terry rating computed with Arena's default style control, so higher means people preferred that model's answers more often — it's a relative rating, not a percentage, so compare it only to other models on this board.
1416.5
1416.5
median 1321.4open leader 1488.2
Overall best: Gemini 4 Carbon (High) 1527.5
80th percentile · #38 of 188 open · #134 of 408 all
#38 of 188
#134 of 408
Arena
Arena · Oct 2026

Score vs VRAM in the ≤96 GB tier

Qwen3.5 122B A10B
Other models
Best Score for its VRAM
x: VRAM (GB, log scale) · y: llmrun Score

#8 of 62 that fit ≤96 GB

Each dot is a model at the best quant that fits ≤96 GB. The line steps up to the best Score available at each amount of VRAM; a dot below it is beaten by a model that needs no more memory.

Full ≤96 GB ranking →

Qwen3.5 122B A10B vs its neighbours

Same boards for the nearest open models per llmrun Score. The highest score in each row is marked.

Board	Qwen3.5 122B A10Bthis model	Qwen3.8 27Bbest in classQ4_K_M · 18 GB	MiniMax M2.5nearest by ScoreQ4_K_M · 140 GB	Qwen3 235B A22B Thinking 2507nearest by ScoreQ4_K_M · 143 GB	MiniMax M2.7nearest by ScoreQ4_K_M · 140 GB
llmrun Score	109.4	122.0	111.8	108.4	107.4
CritPt	0.9%	5.4%	—	0.0%	0.6%
DTBench	84.3%	88.0%	—	80.3%	—
LMCA	32.2%	41.4%	—	29.3%	—
Mystery Game Puzzles	17.0%	—	—	9.0%	—
SciCode	35.6%	46.6%	—	41.4%	50.1%
MMLU-Pro	86.7%	—	80.1%	84.5%	—
LMArena Text	1416.5	1435.6	1390.6	1400.1	1415.0
Compare head to head		Compare	Compare	Compare	Compare

Scores aggregated from public benchmark sources (Epoch AI: 5 boards, Arena: 2 boards, TIGER-Lab: 1 board) — each linked from the benchmark pages. llmrun does not run these benchmarks. Percentiles and ranks are among open models in the llmrun catalogue; the median and leader marks refer to the same field.