Qwen3 Coder 480B A35B Instruct — Benchmarks

Benchmark scores for Qwen3 Coder 480B A35B Instruct 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 →

103.9

90% range 99.5–108.3 · from 6 boards

Coding100.690% range 91–110
Agents104.990% range 100–110
MathNot enough data yet
ScienceNot enough data yet
ReasoningNot enough data yet

Qwen3 Coder 480B A35B Instruct coding benchmarks

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

SWE-bench Verified
SWE-bench Verified tests whether a model can resolve real GitHub issues from popular open-source Python projects, scored on the official swebench.com leaderboard as the percentage of human-validated issues actually fixed. It is the headline measure of practical, agentic software-engineering ability — where open-weight models like Qwen3-Coder, GLM-4.6, Kimi K2 and DeepSWE are now competitive with the frontier.
69.6%
69.6%
median 62.2%open leader 75.8%
Overall best: live-SWE-agent + Claude 4.5 Opus medium (20251101) 79.2%
74th percentile · #6 of 20 open · #47 of 162 all
#6 of 20
#47 of 162
Independent
SWE-bench · Aug 2025
Terminal-Bench
Terminal-Bench measures whether a model can complete real, end-to-end tasks in a command-line environment — running commands, editing files, and chaining steps — making it an agentic test of practical software skill.
27.2%
27.2%
median 31.7%open leader 52.4%
Overall best: GPT 5.5 (unspecified) 84.7%
33rd percentile · #11 of 16 open · #47 of 57 all
#11 of 16
#47 of 57
Lab-reported
Epoch AI · Jul 2025
SWE-bench Bash Only
SWE-bench Bash Only runs the SWE-bench Verified issues through a minimal, single-tool bash agent — no specialised scaffolding — so the score reflects the model's own agentic coding ability rather than the harness around it. A cleaner, harder read on raw software-engineering skill.
55.4%
55.4%
median 55.4%open leader 63.4%
Overall best: Claude 4.5 Opus (high reasoning) 76.8%
56th percentile · #4 of 9 open · #31 of 48 all
#4 of 9
#31 of 48
Independent
SWE-bench · Aug 2025
WeirdML
WeirdML asks models to write machine-learning code for small, deliberately unusual datasets, then runs the code and scores the accuracy it actually achieves. Built by Håvard Tveit Ihle, it rewards practical problem-solving over recognizing a familiar task.
41.2%
41.2%
median 39.5%open leader 82.6%
Overall best: GPT 6 Astra Promax 93.6%
60th percentile · #19 of 46 open · #101 of 162 all
#19 of 46
#101 of 162
Independent
Epoch AI · Jul 2025
ALE-Bench
ALE-Bench, from Sakana AI and AtCoder, tests long-horizon algorithmic optimisation using problems drawn from AtCoder Heuristic Contests, where there is no single correct answer and solutions are scored on quality. Results are reported as an AtCoder-style performance rating, where higher is better.
461.4
461.4
median 694.1open leader 1524.5
Overall best: GPT 6 Astra Max 2951.3
24th percentile · #33 of 43 open · #102 of 122 all
#33 of 43
#102 of 122
Lab-reported
Epoch AI · Jul 2025
GSO-Bench
GSO-Bench, built by a team of UC Berkeley-affiliated researchers, gives a coding agent a real codebase and a performance test and asks it to optimize runtime efficiency to match an expert developer's speedup without breaking correctness, across 102 tasks in 10 codebases and five languages. The scored metric, Opt@1, is the share of tasks where a single attempt reaches at least 95% of the human speedup while still passing all tests.
4.9%
4.9%
median 4.9%open leader 4.9%
Overall best: Claude Fable 5.1 (unspecified) 88.2%
75th percentile · #2 of 3 open · #32 of 42 all
#2 of 3
#32 of 42
Lab-reported
Epoch AI · Jul 2025

Other results (1)· 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.
1274.7
1274.7
median 1411.4open leader 1655.2
Overall best: Claude Opus 5.5 (Max) 1807.3
14th percentile · #45 of 52 open · #120 of 138 all
#45 of 52
#120 of 138
Arena
Arena · Oct 2026

Qwen3 Coder 480B A35B Instruct 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.
1387.6
1387.6
median 1321.5open leader 1487.1
Overall best: Gemini 4 Argon (High) 1523.9
69th percentile · #58 of 187 open · #173 of 405 all
#58 of 187
#173 of 405
Arena
Arena · Oct 2026

Score vs VRAM in the ≤512 GB tier

Qwen3 Coder 480B A35B Instruct
Other models
Best Score for its VRAM
x: VRAM (GB, log scale) · y: llmrun Score

#39 of 96 that fit ≤512 GB

Each dot is a model at the best quant that fits ≤512 GB. The line joins the models with the best Score for their VRAM; a dot below it is beaten by a model that needs no more memory.

Full ≤512 GB ranking →

Qwen3 Coder 480B A35B Instruct vs its neighbours

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

Board	Qwen3 Coder 480B A35B Instructthis model	Qwen3.8 27Bbest in classQ4_K_M · 18 GB	DeepSeek R1 0528nearest by ScoreQ4_K_M · 412 GB	GLM 4.6nearest by ScoreQ4_K_M · 217 GB	GLM 4.5nearest by ScoreQ4_K_M · 218 GB
llmrun Score	103.9	122.1	103.7	103.1	100.9
SWE-bench Verified	69.6%	—	—	68.2%	64.2%
Terminal-Bench	27.2%	—	—	24.5%	—
SWE-bench Bash Only	55.4%	—	—	55.4%	54.2%
WeirdML	41.2%	—	41.6%	—	40.6%
ALE-Bench	461.4	—	804.1	340.8	344.8
LMArena Text	1387.6	1437.8	1421.4	1424.1	1410.8
Compare head to head		Compare	Compare	Compare	Compare

Scores aggregated from public benchmark sources (Epoch AI: 4 boards, Arena: 2 boards, SWE-bench: 2 boards) — 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.