# Qwen3.5 397B vs Spark-X2.6-4B

> Side-by-side benchmark comparison for Qwen3.5 397B and Spark-X2.6-4B across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks.

- Canonical page: https://benchlm.ai/compare/qwen3-5-397b-vs-spark-x2-6-4b
- Last updated: October 20, 2026

- Shared sourced benchmarks: 13
- HTML indexing: indexable
- Ranking lane: BenchAlign v5.8

## Quick Verdict

Qwen3.5 397B leads by point estimate. Conditional score ranges do not establish rank confidence. Choose using the category evidence and your own workload tests.

## Summary

- Qwen3.5 397B has the higher overall point estimate 55.39 to 40.14. Conditional score ranges do not establish rank confidence.
- The clearest category separation is in reasoning, where the averages are 74.1 for Qwen3.5 397B and 48.8 for Spark-X2.6-4B.
- The biggest single benchmark swing is τ³-bench results in Agentic, with scores of 68.4% and 34.7%.
- Spark-X2.6-4B is the cheaper option on output tokens, which matters if you expect large responses or heavy interactive use.
- Spark-X2.6-4B also has the larger context window at 1M.

## Model Snapshot

| Property | Qwen3.5 397B | Spark-X2.6-4B |
|----------|----------|----------|
| Creator | Alibaba | SparkLLM |
| Type | Open Weight | Open Weight |
| Reasoning | Non-Reasoning | Reasoning |
| Context | 128K | 1M |
| Overall Score | 55.39 | 40.14 |
| Benchmarks Covered | 62 | 20 |
| Pricing (input/output) | $0.60 / $3.60 | $0.00 / $0.00 |

## Category Breakdown

### Agentic

- Winner: Directional only
- Qwen3.5 397B public-lane score: 44.2 (Estimated · #62/131)
- Spark-X2.6-4B public-lane score: 33 (Estimated · #89/131)

| Benchmark | Qwen3.5 397B | Spark-X2.6-4B | Winner |
|-----------|-----------|-----------|--------|
| Terminal-Bench 2.0 | 52.5% | Coming soon | Coming soon |
| BrowseComp | 69% | 53.1% | Qwen3.5 397B |
| OSWorld-Verified | 62.2% | Coming soon | Coming soon |
| Claw-Eval | 56.8% | 69.9% | Spark-X2.6-4B |
| QwenClawBench | 51.8% | Coming soon | Coming soon |
| τ³-bench results | 68.4% | 34.7% | Qwen3.5 397B |
| VITA-Bench | 43.7% | Coming soon | Coming soon |
| DeepPlanning | 37.6% | Coming soon | Coming soon |
| Toolathlon | 36.3% | Coming soon | Coming soon |
| MCP Atlas | 46.1% | 62.4% | Spark-X2.6-4B |
| MCP-Tasks | 74.2% | Coming soon | Coming soon |
| WideResearch | 74.0% | Coming soon | Coming soon |
| Gert Labs | 46.76% | Coming soon | Coming soon |
| ResearchClawBench | 14.2% | Coming soon | Coming soon |
| BFCL v4 | Coming soon | 71.8% | Coming soon |
| τ²-bench results | Coming soon | 86.9% | Coming soon |
| HLE w/ tools | Coming soon | 36.9% | Coming soon |

### Coding

- Winner: Not comparable
- Qwen3.5 397B public-lane score: Coming soon
- Spark-X2.6-4B public-lane score: 30.6 (Estimated · #95/152)

| Benchmark | Qwen3.5 397B | Spark-X2.6-4B | Winner |
|-----------|-----------|-----------|--------|
| SWE-bench Verified | 76.2% | 62.8% | Qwen3.5 397B |
| LiveCodeBench v6 | 83.6% | Coming soon | Coming soon |
| SWE-bench Pro | 50.9% | 54.3% | Spark-X2.6-4B |
| SWE Multilingual | 69.3% | 64.7% | Qwen3.5 397B |
| NL2Repo | 32.2% | Coming soon | Coming soon |
| Terminal-Bench 2.0 | 52.5% | Coming soon | Coming soon |
| SciCode | Coming soon | 36.8% | Coming soon |

### Reasoning

- Winner: Not comparable
- Qwen3.5 397B public-lane score: 74.1 (Unranked · 2 rankable rows)
- Spark-X2.6-4B public-lane score: 48.8 (Unranked · 2 rankable rows)

| Benchmark | Qwen3.5 397B | Spark-X2.6-4B | Winner |
|-----------|-----------|-----------|--------|
| LongBench v2 | 63.2% | 53.3% | Qwen3.5 397B |
| AI-Needle | 68.7% | Coming soon | Coming soon |

### Multimodal & Grounded

- Winner: Not comparable
- Qwen3.5 397B public-lane score: 73.6 (#29/55)
- Spark-X2.6-4B public-lane score: Coming soon

| Benchmark | Qwen3.5 397B | Spark-X2.6-4B | Winner |
|-----------|-----------|-----------|--------|
| MMMU | 85.0% | Coming soon | Coming soon |
| MMMU-Pro | 79% | Coming soon | Coming soon |
| RealWorldQA | 83.9% | Coming soon | Coming soon |
| OmniDocBench 1.5 | 90.8% | Coming soon | Coming soon |
| Video-MME (with subtitle) | 87.5% | Coming soon | Coming soon |
| Video-MME (w/o subtitle) | 83.7% | Coming soon | Coming soon |
| MathVision | 88.6% | Coming soon | Coming soon |
| We-Math | 87.9% | Coming soon | Coming soon |
| DynaMath | 86.3% | Coming soon | Coming soon |
| SimpleVQA | 67.1% | Coming soon | Coming soon |
| CharXiv | 80.8% | Coming soon | Coming soon |
| MMLongBench-Doc | 61.5% | Coming soon | Coming soon |
| CC-OCR | 82.0% | Coming soon | Coming soon |
| AI2D_TEST | 93.9% | Coming soon | Coming soon |
| CountBench | 97.2% | Coming soon | Coming soon |
| RefCOCO (avg) | 92.3% | Coming soon | Coming soon |
| ODINW13 | 47.0% | Coming soon | Coming soon |
| ERQA | 67.5% | Coming soon | Coming soon |
| VideoMMMU | 84.7% | Coming soon | Coming soon |
| MLVU (M-Avg) | 86.7% | Coming soon | Coming soon |
| ScreenSpot Pro | 65.6% | Coming soon | Coming soon |
| V* | 95.8% | Coming soon | Coming soon |

### Knowledge

- Winner: Directional only
- Qwen3.5 397B public-lane score: 53.2 (Estimated · #70/187)
- Spark-X2.6-4B public-lane score: 43 (Estimated · #109/187)

| Benchmark | Qwen3.5 397B | Spark-X2.6-4B | Winner |
|-----------|-----------|-----------|--------|
| GPQA | 88.4% | 69.9% | Qwen3.5 397B |
| SuperGPQA | 70.4% | Coming soon | Coming soon |
| MMLU-Pro | 87.8% | Coming soon | Coming soon |
| MMLU-Redux | 94.9% | Coming soon | Coming soon |
| C-Eval | 93% | Coming soon | Coming soon |
| HLE | 28.7% | Coming soon | Coming soon |
| GPQA-D | Coming soon | 77.1% | Coming soon |
| HLE w/o tools | Coming soon | 17.9% | Coming soon |

### Multilingual

- Winner: Not comparable
- Qwen3.5 397B public-lane score: 93.8 (#5/17)
- Spark-X2.6-4B public-lane score: Coming soon

| Benchmark | Qwen3.5 397B | Spark-X2.6-4B | Winner |
|-----------|-----------|-----------|--------|
| MMLU-ProX | 84.7% | Coming soon | Coming soon |
| NOVA-63 | 59.1% | Coming soon | Coming soon |
| INCLUDE | 85.6% | Coming soon | Coming soon |
| PolyMath | 73.3% | Coming soon | Coming soon |
| MAXIFE | 88.2% | Coming soon | Coming soon |

### Instruction Following

- Winner: Qwen3.5 397B
- Qwen3.5 397B public-lane score: 79.9 (#59/149)
- Spark-X2.6-4B public-lane score: 78.2 (#62/149)

| Benchmark | Qwen3.5 397B | Spark-X2.6-4B | Winner |
|-----------|-----------|-----------|--------|
| IFEval | 92.6% | 93.6% | Spark-X2.6-4B |
| IFBench | 76.5% | 75.8% | Qwen3.5 397B |

### Mathematics

- Winner: Not comparable
- Qwen3.5 397B public-lane score: 81.7 (Unranked · 5 rankable rows)
- Spark-X2.6-4B public-lane score: 76.4 (Unranked · 3 rankable rows)

| Benchmark | Qwen3.5 397B | Spark-X2.6-4B | Winner |
|-----------|-----------|-----------|--------|
| AIME26 | 93.3% | 91.2% | Qwen3.5 397B |
| HMMT Feb 2025 | 94.8% | Coming soon | Coming soon |
| HMMT Nov 2025 | 92.7% | Coming soon | Coming soon |
| HMMT Feb 2026 | 87.9% | 81.3% | Qwen3.5 397B |
| MMAnswerBench | 80.9% | Coming soon | Coming soon |
| IMOAnswerBench | Coming soon | 75.2% | Coming soon |

## FAQ

### Which is better overall, Qwen3.5 397B or Spark-X2.6-4B?

Qwen3.5 397B has the higher overall point estimate. Conditional score ranges do not establish rank confidence.

### Where is the biggest gap between Qwen3.5 397B and Spark-X2.6-4B?

The widest category gap is in reasoning, where the averages are 74.1 for Qwen3.5 397B and 48.8 for Spark-X2.6-4B.

### How many benchmarks does Qwen3.5 397B cover on BenchLM?

Qwen3.5 397B currently has 62 sourced benchmark scores on BenchLM.

### How many benchmarks does Spark-X2.6-4B cover on BenchLM?

Spark-X2.6-4B currently has 20 sourced benchmark scores on BenchLM.

## Related Comparisons

- [Qwen3.5 397B vs Qwen3.5 397B (Reasoning)](/compare/qwen3-5-397b-vs-qwen3-5-397b-reasoning)
- [Spark-X2.6-4B vs Qwen3.5 397B (Reasoning)](/compare/qwen3-5-397b-reasoning-vs-spark-x2-6-4b)
- [Qwen3.5 397B vs Claude Opus 5.5](/compare/claude-opus-5-5-vs-qwen3-5-397b)
- [Spark-X2.6-4B vs Claude Opus 5.5](/compare/claude-opus-5-5-vs-spark-x2-6-4b)

## Explore More

- [Qwen3.5 397B profile](/models/qwen3-5-397b)
- [Spark-X2.6-4B profile](/models/spark-x2-6-4b)
- [Compare Pricing](/llm-pricing)
- [Alternative Finder](/tools/alternative-finder)
- [LLM Selector](/tools/llm-selector)
- [Overall Rankings](/best/overall)