Qwen3 Coder 480B A35B qwen3-coder-480b-a35b
qwen · qwen coding flagship
#60 overall #7 agentic
- params
- 480B (A35B)
- arch
- moe
- context
- 256k
- license
- apache-2.0 open weights
- released
- Jul 2025
- train compute
- 1.6e+24 FLOP
# architecture
- layers
- 62
- d_model
- 6144
- heads
- 96q / 8kv
- head dim
- 128
- experts
- top-8 of 160
- vocab
- 151936
- family
- qwen3_moe
signal path of one layer, generated from the registry's structured fields — dimension lines quote the model's real numbers; an MoE trace forks at the router, a dense trace runs straight through. Geometry fields come from the repo's config.json.
# why ranked
Overall: #60 — score 43.4 ● high — all signals present
| signal | weight | input (0–100) |
|---|---|---|
| aa_intelligence | 0.70 | 43.1 |
| bench_composite | 0.30 | 44.2 |
benchmark panel evidence:
| complete panel | score (0–100) |
|---|---|
| aa-math-v1 | 18.9 |
| swe-bench-verified-v1 | 69.6 |
Coding: provisional — score 73.7 ● low — a single signal; treat with caution
provisional: too few independent signals for a numbered position — the score is shown, but this model sorts after every ranked model.
| signal | weight | input (0–100) |
|---|---|---|
| aa_coding | 0.30 | — |
| aider_polyglot | 0.30 | — |
| swe_bench_verified | 0.40 | 73.7 |
missing signals are dropped and the remaining weights renormalized — never imputed.
Agentic: #7 — score 73.7 ● low — a single signal; treat with caution
| signal | weight | input (0–100) |
|---|---|---|
| swe_bench_verified | 0.60 | 73.7 |
| swe_rebench | 0.40 | — |
missing signals are dropped and the remaining weights renormalized — never imputed.
# trend
methodology changed during this history window; score movement across that boundary is not model movement. See methodology v6.
# benchmarks
| benchmark | score | source | date |
|---|---|---|---|
| AA Intelligence Index | 11.9 | Artificial Analysis | — |
| AA Math Index | 39.3 | Artificial Analysis | — |
| SWE-bench Verified | 69.6 | SWE-bench | Aug 2025 |
# where to run
| provider | quant | ctx | $/M in | $/M out | $/M cache | price src | tps | uptime | ✓ |
|---|---|---|---|---|---|---|---|---|---|
| SiliconFlow | 262k | $0.25 | $1.00 | — | via modelsdev | — | — | ||
| DeepInfra | fp4 | 256k | $0.30 | $1.00 | $0.10 | via openrouter | — | 85.1% | |
| unknown | 256k | $0.22 | $1.80 | — | via openrouter | — | — | ||
| Venice | fp8 | 250k | $0.35 | $1.50 | $0.04 | via openrouter | — | 75.5% | |
| Novita | 256k | $0.38 | $1.55 | — | via litellm | — | — | ||
| Amazon Bedrock | — | $0.45 | $1.80 | — | bedrock | — | — | ✓ | |
| Hyperbolic | 256k | $2.00 | $2.00 | — | hyperbolic | — | — | ✓ | |
| Together | 256k | $2.00 | $2.00 | — | via modelsdev | — | — | ||
| Alibaba | 256k | $1.50 | $7.50 | — | via modelsdev | — | — |
sorted by blended price ((3·input + output) / 4 per 1M) · ✓ = the provider's own catalog confirms the offer · "via …" prices are what the aggregator routing the offer charges, not the provider's own list price
source aliases
- aa
qwen3-coder-480b,qwen3-coder-480b-a35b-instruct- arena
Qwen3-Coder-480B-A35B-Instruct- openrouter
qwen/qwen3-coder- swebench
Qwen3-Coder-480B-A35B-Instruct