LLMs TigerLive LLM benchmarks

Qwen3-Coder-480B-A35B-Instruct: benchmark scores, price and ranking

Qwen3-Coder-480B-A35B-Instruct from Alibaba ranks #581 of 743 LLMs on the LLMs Tiger aggregate benchmark score (45.5), based on 6 benchmark results captured up to 2026-10-13.

Organisation
Alibaba
Type
Open-weight / local model
Open weights
Yes
Released
2025-07-22 (source: curated)
Cheapest API price
$0.380 in / $1.55 out per million tokens (via novita, observed 2026-10-13T20:47Z)
Aggregate rank
#581 of 743 (score 45.5)

Relative to other models, Qwen3-Coder-480B-A35B-Instruct is strongest on WeirdML (#55 of 109) and weakest on WebDev Arena (#56 of 67).

Across the 6 benchmarks where at least ten models are scored, Qwen3-Coder-480B-A35B-Instruct is in the top 10% on 0 and in the top half on 0.

Its cheapest listed input price, $0.380 per million tokens, is lower than that of 57% of the 427 models with a listed price.

Qwen3-Coder-480B-A35B-Instruct benchmark results

Qwen3-Coder-480B-A35B-Instruct benchmark scores, ranked against all models measured on each benchmark
BenchmarkScoreRankSourceCaptured
WeirdML41.2%#55 of 109Epoch AI Benchmarking Hub2026-10-13
SWE-bench Verified55.4%#39 of 73SWE-bench Verified leaderboard2026-10-13
GSO-Bench4.9%#11 of 19Epoch AI Benchmarking Hub2026-10-13
ALE-Bench461#60 of 79Epoch AI Benchmarking Hub2026-10-13
Terminal-Bench 2.027.2%#22 of 28Epoch AI Benchmarking Hub2026-10-13
WebDev Arena1,274#56 of 67Epoch AI Benchmarking Hub2026-10-13

About Qwen3-Coder-480B-A35B-Instruct

Qwen3-Coder-480B-A35B-Instruct is an open-weight model that can run on your own hardware from Alibaba. It was released on 2025-07-22, according to the developer's announcement. The cheapest API listing we track charges $0.380 per million input tokens and $1.55 per million output tokens, through novita. We hold 6 benchmark results for it across 2 areas, and none is in the top tenth of its board.

Where Qwen3-Coder-480B-A35B-Instruct stands

On our aggregate score, Qwen3-Coder-480B-A35B-Instruct is #581 of 743 ranked models, in the lower half. Its score is 45.5; the scale is centred on 50, so it sits 4.5 points below the typical model. That beats 22% of the ranked field. Among the 74 ranked Alibaba models it is #63. The next one up is qwen1.5-72b-chat at 45.5. The next one down is Qwen2.5-Coder-1.5B-Instruct at 45.4.

Qwen3-Coder-480B-A35B-Instruct, benchmark by benchmark

Qwen3-Coder-480B-A35B-Instruct lands below the middle on WeirdML: 41.2%, #55 of 109, with 50% of the field scoring higher. GPT-6 Astra (max) leads at 93.3%; the closest result above is DeepSeek-R1 (May 2025) at 41.6%. It covers unusual machine-learning tasks. The median score on this board is also 41.2%. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Qwen3-Coder-480B-A35B-Instruct lands below the middle on SWE-bench Verified: 55.4%, #39 of 73, with 52% of the field scoring higher. Claude Opus 4.7 (max) leads at 83.5%; the closest result above is GLM-4.6 at 55.4%. It covers real GitHub issues resolved. The board median is 56.4%, so Qwen3-Coder-480B-A35B-Instruct is 1.0 points below it. Source: SWE-bench Verified leaderboard, captured 2026-10-13.

Qwen3-Coder-480B-A35B-Instruct lands below the middle on GSO-Bench: 4.9%, #11 of 19, with 53% of the field scoring higher. GPT-5.5 (xhigh) leads at 40.2%; the closest result above is Kimi K2 Instruct at 4.9%. It covers software performance optimisation. The median score on this board is also 4.9%. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Qwen3-Coder-480B-A35B-Instruct lands below the middle on ALE-Bench: 461, #60 of 79, with 75% of the field scoring higher. GPT-6 Astra (max) leads at 2,951; the closest result above is Mistral Small 4 at 498. It covers AtCoder heuristic-optimisation contests (performance rating). The board median is 764, so Qwen3-Coder-480B-A35B-Instruct is 303 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Qwen3-Coder-480B-A35B-Instruct lands below the middle on Terminal-Bench 2.0: 27.2%, #22 of 28, with 75% of the field scoring higher. Gemini 3 Pro Preview leads at 69.4%; the closest result above is grok-4-0709 at 27.2%. It covers real terminal tasks (agent + model). The board median is 37.3%, so Qwen3-Coder-480B-A35B-Instruct is 10.1 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

WebDev Arena is a weak spot. Qwen3-Coder-480B-A35B-Instruct scores 1,274, #56 of 67, behind 82% of the models measured. The leader, Claude Opus 5.5 (max), reaches 1,820. It covers building web apps, crowd-voted (Elo). The board median is 1,433, so Qwen3-Coder-480B-A35B-Instruct is 160 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Strengths and weaknesses

Software engineering: uneven, ahead of 30% of the field on average across 5 benchmarks. Its best board here is SWE-bench Verified (#39 of 73) and its weakest is WebDev Arena (#56 of 67).

Price and value

The median input price across the 427 priced models is $0.500 per million tokens, so Qwen3-Coder-480B-A35B-Instruct costs less than the typical listing. Models that score within 1.5 points or better and cost less: Qwen2.5-Coder-7B-Instruct ($0.010, 49.5), DeepSeek V4.1 Flash (high) ($0.010, 48.9), Qwen2.5-Coder-3B-Instruct ($0.010, 46.5). Similar scores cost more on qwen-plus-2025-01-25 ($0.400) and mistral-medium-2505 ($0.400).

Release history

The Alibaba model released before it was qwen3-4b on 2025-04-29, which scores 47.9. The next one was Qwen3-235B-A22B-Instruct-2507 on 2025-07-25, at 49.3.

Reading these numbers

The 6 scores on this page come from 2 sources: Epoch AI Benchmarking Hub (5), SWE-bench Verified leaderboard (1). 1 of them come from an official leaderboard and 5 from datasets or third-party runs. With 6 results the aggregate rank can move a lot when one new score arrives.

Models ranked nearby

More from Alibaba

Compare it on the live leaderboard → · How the aggregate score works