# Grok 4.7 vs Qwen3 8B

> Grok 4.7 is the stronger model overall, scoring 46.6 to 33.3 on the Noometry Index. Qwen3 8B costs 9.7× less per token, which makes it the better buy when Grok 4.7's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/grok-4-7-vs-qwen3-8b
- Last updated: 2026-10-10
- Shared benchmarks: 2

## Summary

- They share 2 benchmarks with published results for both. Grok 4.7 scores higher in 5 categories and Qwen3 8B in 0 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Grok 4.7 leads 40.1 to 16.1.
- The biggest single-benchmark swing is SciCode: 57.4% for Grok 4.7 and 22.6% for Qwen3 8B.
- Qwen3 8B is cheaper at $0.18 / $0.70 per million input/output tokens, against $2 / $6 for Grok 4.7.
- Grok 4.7 accepts more context: 500K tokens versus 131K.
- Qwen3 8B has downloadable open weights; the other is API-only.

## Snapshot

| | Grok 4.7 | Qwen3 8B |
|---|---|---|
| Provider | xAI | Alibaba (Qwen) |
| Noometry Index | 46.6 | 33.3 |
| Rank | 56 | 237 |
| Context | 500K | 131K |
| Input $/M | $2 | $0.18 |
| Output $/M | $6 | $0.70 |
| Weights | Proprietary | Open |

## Coding

- Grok 4.7: 55.0 (#23)
- Qwen3 8B: 33.4 (#245)

| Benchmark | Grok 4.7 | Qwen3 8B |
|---|---|---|
| SciCode | 57.4% | 22.6% |
| CursorBench | 46.3% | — |
| LMArena WebDev | 1636 | — |
| FrontierSWE | 29.5% | — |
| LMArena Coding | 1423 | — |

## Agentic & Tool Use

- Grok 4.7: —
- Qwen3 8B: 29.9 (#71)

| Benchmark | Grok 4.7 | Qwen3 8B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 42.6% |

## Reasoning

- Grok 4.7: 40.1 (#49)
- Qwen3 8B: 16.1 (#299)

| Benchmark | Grok 4.7 | Qwen3 8B |
|---|---|---|
| CritPt | 17.7% | 0% |
| NYT Connections (extended) | 76.8% | — |
| Chess Puzzles | — | 5% |
| LMArena Hard Prompts | 1409 | — |
| DTBench | — | 59.7% |
| LMCA | — | 8.8% |
| Epoch Capabilities Index | — | 136.17 |

## Math

- Grok 4.7: 41.1 (#73)
- Qwen3 8B: 33.3 (#176)

| Benchmark | Grok 4.7 | Qwen3 8B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | — | 56.1% |
| ProofBench | 34% | — |
| LMArena Math | 1393 | — |

## Knowledge

- Grok 4.7: 38.7 (#130)
- Qwen3 8B: 36.1 (#164)

| Benchmark | Grok 4.7 | Qwen3 8B |
|---|---|---|
| GPQA Diamond | — | 56.8% |
| Vectara Hallucination Rate | — | 4.8% |
| LMArena Expert | 1413 | — |

## Multimodal

- Grok 4.7: 28.8 (#107)
- Qwen3 8B: —

| Benchmark | Grok 4.7 | Qwen3 8B |
|---|---|---|
| Furniture Assembly | 20.8% | — |

## Multilingual

- Grok 4.7: 51.2 (#108)
- Qwen3 8B: —

| Benchmark | Grok 4.7 | Qwen3 8B |
|---|---|---|
| LMArena Non-English | 1392 | — |
| LMArena Chinese | 1476 | — |
| LMArena Russian | 1396 | — |

## Instruction Following

- Grok 4.7: 74.3 (#98)
- Qwen3 8B: —

| Benchmark | Grok 4.7 | Qwen3 8B |
|---|---|---|
| LMArena Instruction Following | 1406 | — |

## Long Context

- Grok 4.7: 43.2 (#97)
- Qwen3 8B: 37.8 (#205)

| Benchmark | Grok 4.7 | Qwen3 8B |
|---|---|---|
| Fiction.LiveBench | — | 62.1% |
| LMArena Longer Query | 1416 | — |

## Writing & Preference

- Grok 4.7: 70.1 (#23)
- Qwen3 8B: —

| Benchmark | Grok 4.7 | Qwen3 8B |
|---|---|---|
| LMArena Text | 1397 | — |
| LMArena Creative Writing | 1391 | — |
| EQ-Bench Creative Writing | 2007 | — |
| LMArena Multi-Turn | 1389 | — |

## FAQ

### Is Grok 4.7 better than Qwen3 8B?

Grok 4.7 is the stronger model overall, scoring 46.6 to 33.3 on the Noometry Index. Qwen3 8B costs 9.7× less per token, which makes it the better buy when Grok 4.7's lead doesn't matter for your workload.

### Which is cheaper, Grok 4.7 or Qwen3 8B?

Qwen3 8B is cheaper. It lists at $0.18 per million input tokens and $0.70 per million output tokens; Grok 4.7 lists at $2 and $6.

### Is Grok 4.7 or Qwen3 8B better for coding?

Grok 4.7 scores higher on coding benchmarks: 55.0 versus 33.4 in the Noometry coding category.

### Which has the bigger context window?

Grok 4.7 does, with 500K tokens against 131K.

### How many benchmarks do Grok 4.7 and Qwen3 8B share?

2 benchmarks have published results for both models. Grok 4.7 has 21 scored results on Noometry and Qwen3 8B has 11.