Qwen3.8 27B
| Released | 2026-08-14 |
|---|---|
| Price per 1M | $0.20 in · $2.70 out |
| Overall score | 56.0 · rank 68 |
| Context window | 262,144 tokens |
| Provider | Alibaba |
| Weights | Open weights |
Scores
| Dimension | Score | Rank | Benchmarks |
|---|---|---|---|
| Overall: Balanced | 56.0 | 68 | 35 |
| Overall: Coding | 57.1 | 58 | 26 |
| Overall: Reasoning | 55.2 | 72 | 11 |
| Overall: Agentic work | 56.8 | 62 | 23 |
| Overall: Frontend | 56.5 | 52 | 11 |
| Coding | 57.7 | 47 | 6 |
| Agents | 56.3 | 65 | 9 |
| Reasoning | 57.0 | 68 | 8 |
| Math | 54.4 | 77 | 1 |
| Knowledge | 52.6 | 80 | 2 |
| Frontend | 56.0 | 47 | 3 |
| General | 54.4 | 70 | 4 |
| Multimodal | 55.6 | 44 | 2 |
| Value vs price | +8.8 | 40 |
Configurations
| Configuration | Balanced | Benchmarks |
|---|---|---|
| Unspecified setting | 57.0 ±1.3 | 17 |
| Low | 47.9 ±5.2 | 4 |
| Medium | 47.0 ±5.2 | 4 |
| Non-reasoning | 43.6 ±4.9 | 4 |
| Extra High | 55.7 ±1.2 | 19 |
Evidence
Coding 57.7 ±2.6
| Benchmark | Result | Source | Date |
|---|---|---|---|
| LiveBench 2.0 Coding | 69.3% | LiveBench | 2026-11-06 |
| IOI | 39.1% | Vals AI | 2026-10-29 |
| SWE-bench Verified (Vals run) | 86.0% | Vals AI | 2026-09-01 |
| SciCode | 46.6% | Epoch AI | 2026-11-11 |
| LiveCodeBench (Vals run) | 84.0% | Vals AI | 2026-09-01 |
| ProgramBench | 0.0% | Vals AI | 2026-11-09 |
Agents 56.3 ±1.4
| Benchmark | Result | Source | Date |
|---|---|---|---|
| Terminal-Bench-Science | 0.0% | Vals AI | 2026-11-09 |
| APEX-Agents | 47.5% | Epoch AI | 2026-11-11 |
| LiveBench 2.0 Agentic | 15.9% | LiveBench | 2026-11-06 |
| LiveBench 2.0 Browser Use | 24.2% | LiveBench | 2026-11-06 |
| Terminal-Bench 2.1 (Vals run) | 58.4% | Vals AI | 2026-10-01 |
| LiveBench 2.0 Research | 61.3% | LiveBench | 2026-11-06 |
| Tax Agent Bench | 30.9% | Vals AI | 2026-11-09 |
| SkillsBench | 38.1% | Vals AI | 2026-10-29 |
| CyberBench v1.1 | 67.8% | Vals AI | 2026-11-05 |
Reasoning 57.0 ±1.6
| Benchmark | Result | Source | Date |
|---|---|---|---|
| Extended NYT Connections | 54.5% | Lech Mazur | 2026-11-10 |
| ARC-AGI-2 | 42.4% | Epoch AI | 2026-11-11 |
| ARC-AGI-2 (ARC Prize verified) | 42.4% | ARC Prize | 2026-11-11 |
| GPQA Diamond (Vals run) | 88.9% | Vals AI | 2026-09-01 |
| CritPt | 5.4% | Epoch AI | 2026-11-11 |
| ARC-AGI-1 (ARC Prize verified) | 87.5% | ARC Prize | 2026-11-11 |
| ARC-AGI-1 | 87.5% | Epoch AI | 2026-11-11 |
| SimpleBench | 60.2% | SimpleBench | 2026-08-20 |
| LiveBench 2.0 Reasoning | 80.0% | LiveBench | 2026-11-06 |
| LiveBench 2.0 Data Analysis | 76.6% | LiveBench | 2026-11-06 |
Math 54.4 ±2.9
| Benchmark | Result | Source | Date |
|---|---|---|---|
| ProofBench v1.1 | 16.0% | Vals AI | 2026-11-09 |
| ProofBench | 16.0% | Epoch AI | 2026-11-11 |
Knowledge 52.6 ±3.9
| Benchmark | Result | Source | Date |
|---|---|---|---|
| MMLU-Pro (Vals run) | 84.3% | Vals AI | 2026-09-01 |
| Legal Research Bench | 36.1% | Vals AI | 2026-11-09 |
Frontend 56.0 ±1.7
| Benchmark | Result | Source | Date |
|---|---|---|---|
| LiveBench 2.0 Design | 72.4% | LiveBench | 2026-11-06 |
| Vibe Code Bench v1.1 | 64.9% | Vals AI | 2026-11-09 |
| LMArena WebDev | 1,591 | LMArena | 2026-11-09 |
General 54.4 ±3.4
| Benchmark | Result | Source | Date |
|---|---|---|---|
| EQ-Bench Creative Writing v3 | 1,671 | EQ-Bench | 2026-11-11 |
| LMArena Text | 1,433 | LMArena | 2026-11-09 |
| Creative Story-Writing (Lech Mazur) | 39.0% | Lech Mazur | 2026-10-23 |
| EQ-Bench Longform Creative Writing | 53.3% | EQ-Bench | 2026-11-11 |
Multimodal 55.6 ±2.3
| Benchmark | Result | Source | Date |
|---|---|---|---|
| MMMU-Pro (Vals run) | 83.9% | Vals AI | 2026-09-01 |
| LMArena Vision | 1,243 | LMArena | 2026-11-09 |
Results not used in the score
| Benchmark | Result | Source | Why not used |
|---|---|---|---|
| ARC-AGI-1 | 69.2% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-1 | 68.7% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-1 | 34.0% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-1 (ARC Prize verified) | 69.2% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-1 (ARC Prize verified) | 68.7% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-1 (ARC Prize verified) | 34.0% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-2 | 22.8% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-2 | 13.2% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-2 | 1.5% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-2 (ARC Prize verified) | 22.8% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-2 (ARC Prize verified) | 13.2% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-2 (ARC Prize verified) | 1.5% | ARC Prize | A stronger configuration result was used |
| CritPt | 0.0% | Epoch AI | A stronger configuration result was used |
| CritPt | 0.0% | Epoch AI | A stronger configuration result was used |
| CritPt | 0.3% | Epoch AI | A stronger configuration result was used |
| SciCode | 40.1% | Epoch AI | A stronger configuration result was used |
| SciCode | 39.0% | Epoch AI | A stronger configuration result was used |
| SciCode | 36.2% | Epoch AI | A stronger configuration result was used |
Names matched to this model
qwen-3-8-27b, qwen3-8-27b