Model comparison
DeepSeek-V3.1-Terminus vs Qwen3.8 27B
Qwen3.8 27B is the stronger model overall, scoring 46.0 to 43.1 on the Noometry Index. DeepSeek-V3.1-Terminus costs 2.5× less per token, which makes it the better buy when Qwen3.8 27B's lead doesn't matter for your workload.
Last verified . 14 shared benchmarks.
DeepSeek
Alibaba (Qwen)Summary
- They share 14 benchmarks with published results for both. DeepSeek-V3.1-Terminus scores higher in 1 category and Qwen3.8 27B in 6 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Qwen3.8 27B leads 41.0 to 26.4.
- The biggest single-benchmark swing is LMCA: 28.6% for DeepSeek-V3.1-Terminus and 41.4% for Qwen3.8 27B.
- DeepSeek-V3.1-Terminus is cheaper at $0.27 / $1 per million input/output tokens, against $0.99 / $1.49 for Qwen3.8 27B.
- Qwen3.8 27B accepts more context: 262K tokens versus 164K.
Side by side
| DeepSeek-V3.1-Terminus | Qwen3.8 27B | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 43.1 | 46.0 |
| Released | 2025-09-22 | 2026-08-14 |
| Weights | Open | Open |
| Context window | 164K | 262K |
| Max output | 147K | 33K |
| Input $ / M tokens | $0.27 | $0.99 |
| Output $ / M tokens | $1 | $1.49 |
| Results tracked | 16 | 31 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.8 27B leads
DeepSeek-V3.1-Terminus: 42.0 (#113), Qwen3.8 27B: 50.5 (#44)
| Benchmark | DeepSeek-V3.1-Terminus | Qwen3.8 27B |
|---|---|---|
| SciCode | 40.6% | 46.6% |
| LMArena Coding | 1426 | 1482 |
| LMArena WebDev | — | 1593 |
| ALE-Bench | 745.17 | — |
Agentic & Tool Use Not comparable
DeepSeek-V3.1-Terminus: —, Qwen3.8 27B: 32.9 (#57)
| Benchmark | DeepSeek-V3.1-Terminus | Qwen3.8 27B |
|---|---|---|
| APEX-Agents | — | 47.5% |
Reasoning Qwen3.8 27B leads
DeepSeek-V3.1-Terminus: 26.4 (#133), Qwen3.8 27B: 41.0 (#54)
| Benchmark | DeepSeek-V3.1-Terminus | Qwen3.8 27B |
|---|---|---|
| CritPt | 1.7% | 5.4% |
| LMArena Hard Prompts | 1426 | 1460 |
| DTBench | 81.3% | 88% |
| LMCA | 28.6% | 41.4% |
| ARC-AGI-2 | — | 42.4% |
| Kagi LLM Benchmark | 57.4% | — |
| NYT Connections (extended) | — | 54.5% |
| ARC-AGI-1 | — | 87.5% |
| Surface Evolver Bench | — | 45% |
| Epoch Capabilities Index | — | 149.38 |
Math DeepSeek-V3.1-Terminus leads
DeepSeek-V3.1-Terminus: 38.5 (#137), Qwen3.8 27B: 37.1 (#161)
| Benchmark | DeepSeek-V3.1-Terminus | Qwen3.8 27B |
|---|---|---|
| LMArena Math | 1402 | 1456 |
| ProofBench | — | 16% |
Knowledge Not comparable
DeepSeek-V3.1-Terminus: —, Qwen3.8 27B: 41.6 (#109)
| Benchmark | DeepSeek-V3.1-Terminus | Qwen3.8 27B |
|---|---|---|
| LMArena Expert | — | 1482 |
Multimodal Not comparable
DeepSeek-V3.1-Terminus: —, Qwen3.8 27B: 41.3 (#37)
| Benchmark | DeepSeek-V3.1-Terminus | Qwen3.8 27B |
|---|---|---|
| LMArena Vision | — | 1271 |
Multilingual Qwen3.8 27B leads
DeepSeek-V3.1-Terminus: 52.1 (#92), Qwen3.8 27B: 53.7 (#60)
| Benchmark | DeepSeek-V3.1-Terminus | Qwen3.8 27B |
|---|---|---|
| LMArena Non-English | 1407 | 1430 |
| LMArena Russian | 1436 | 1415 |
| LMArena Chinese | — | 1504 |
| LMArena French | — | 1465 |
| LMArena German | — | 1438 |
| LMArena Japanese | — | 1384 |
| LMArena Korean | — | 1393 |
| LMArena Spanish | — | 1448 |
Instruction Following Qwen3.8 27B leads
DeepSeek-V3.1-Terminus: 74.0 (#106), Qwen3.8 27B: 75.8 (#53)
| Benchmark | DeepSeek-V3.1-Terminus | Qwen3.8 27B |
|---|---|---|
| LMArena Instruction Following | 1404 | 1439 |
Long Context Too close to call
DeepSeek-V3.1-Terminus: 43.4 (#97), Qwen3.8 27B: 44.3 (#70)
| Benchmark | DeepSeek-V3.1-Terminus | Qwen3.8 27B |
|---|---|---|
| LMArena Longer Query | 1421 | 1450 |
Writing & Preference Qwen3.8 27B leads
DeepSeek-V3.1-Terminus: 61.0 (#92), Qwen3.8 27B: 65.8 (#43)
| Benchmark | DeepSeek-V3.1-Terminus | Qwen3.8 27B |
|---|---|---|
| LMArena Text | 1419 | 1441 |
| LMArena Creative Writing | 1403 | 1384 |
| LMArena Multi-Turn | 1411 | 1441 |
| EQ-Bench Creative Writing | — | 1671 |
Frequently asked questions
Is DeepSeek-V3.1-Terminus better than Qwen3.8 27B?
Qwen3.8 27B is the stronger model overall, scoring 46.0 to 43.1 on the Noometry Index. DeepSeek-V3.1-Terminus costs 2.5× less per token, which makes it the better buy when Qwen3.8 27B's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-V3.1-Terminus or Qwen3.8 27B?
DeepSeek-V3.1-Terminus is cheaper. It lists at $0.27 per million input tokens and $1 per million output tokens; Qwen3.8 27B lists at $0.99 and $1.49.
Is DeepSeek-V3.1-Terminus or Qwen3.8 27B better for coding?
Qwen3.8 27B scores higher on coding benchmarks: 50.5 versus 42.0 in the Noometry coding category.
Which has the bigger context window?
Qwen3.8 27B does, with 262K tokens against 164K.
How many benchmarks do DeepSeek-V3.1-Terminus and Qwen3.8 27B share?
14 benchmarks have published results for both models. DeepSeek-V3.1-Terminus has 16 scored results on Noometry and Qwen3.8 27B has 31.