# Claude Haiku 4.5 vs Grok 4.7

> Grok 4.7 is the stronger model overall, scoring 53.1 to 39.5 on the Noometry Index. Claude Haiku 4.5 costs 1.5× 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/claude-haiku-4-5-vs-grok-4-7
- Last updated: 2026-10-10
- Shared benchmarks: 27

## Summary

- They share 27 benchmarks with published results for both. Claude Haiku 4.5 scores higher in 1 category and Grok 4.7 in 9 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Grok 4.7 leads 49.1 to 15.1.
- The biggest single-benchmark swing is NYT Connections (extended): 14.3% for Claude Haiku 4.5 and 76.8% for Grok 4.7.
- Claude Haiku 4.5 is cheaper at $1 / $5 per million input/output tokens, against $2 / $6 for Grok 4.7.
- Grok 4.7 accepts more context: 500K tokens versus 200K.

## Snapshot

| | Claude Haiku 4.5 | Grok 4.7 |
|---|---|---|
| Provider | Anthropic | xAI |
| Noometry Index | 39.5 | 53.1 |
| Rank | 165 | 37 |
| Context | 200K | 500K |
| Input $/M | $1 | $2 |
| Output $/M | $5 | $6 |
| Weights | Proprietary | Proprietary |

## Coding

- Claude Haiku 4.5: 44.0 (#78)
- Grok 4.7: 58.0 (#18)

| Benchmark | Claude Haiku 4.5 | Grok 4.7 |
|---|---|---|
| LMArena WebDev | 1330 | 1639 |
| SciCode | 43.3% | 57.8% |
| LMArena Coding | 1453 | 1427 |
| FrontierCode | — | 47.6% |
| SWE-bench Verified (bash only) | 66.6% | — |
| CursorBench | — | 46.3% |
| SWE-bench Multilingual | 64.7% | — |
| FrontierSWE | — | 29.5% |
| WeirdML | 45.4% | — |
| ALE-Bench | 653.48 | — |

## Agentic & Tool Use

- Claude Haiku 4.5: 33.6 (#52)
- Grok 4.7: 36.7 (#37)

| Benchmark | Claude Haiku 4.5 | Grok 4.7 |
|---|---|---|
| Vending-Bench 2 | 458.89 | 10,537 |
| Terminal-Bench | 35.5% | — |
| APEX-Agents | — | 54.6% |
| Berkeley Function Calling Leaderboard | 68.7% | — |
| DeepResearch Bench | 45.5% | — |
| BALROG | 31.2% | — |
| ExploitBench | 13.7% | — |
| GDP.pdf | — | 22.8% |

## Reasoning

- Claude Haiku 4.5: 15.1 (#320)
- Grok 4.7: 49.1 (#40)

| Benchmark | Claude Haiku 4.5 | Grok 4.7 |
|---|---|---|
| NYT Connections (extended) | 14.3% | 76.8% |
| CritPt | 0% | 18% |
| Chess Puzzles | 8% | 38% |
| LMArena Hard Prompts | 1420 | 1413 |
| DTBench | 73.6% | 96% |
| LMCA | 30.9% | 49.4% |
| Epoch Capabilities Index | 142.41 | 153.53 |
| ARC-AGI-2 | 4% | — |
| ARC-AGI-1 | 47.7% | — |
| Mystery Game Puzzles | — | 29% |
| ForecastBench | 61.4 | — |

## Math

- Claude Haiku 4.5: 44.9 (#78)
- Grok 4.7: 57.8 (#39)

| Benchmark | Claude Haiku 4.5 | Grok 4.7 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.7% | 98.1% |
| LMArena Math | 1396 | 1407 |
| FrontierMath (Tiers 1-3) | — | 53% |
| FrontierMath Tier 4 | — | 17.1% |
| ProofBench | — | 34% |
| Omni-MATH | 56.1% | — |
| MATH Level 5 | 96.4% | — |
| FrontierMath (Feb 2025 set) | 5.9% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |

## Knowledge

- Claude Haiku 4.5: 37.7 (#153)
- Grok 4.7: 62.8 (#22)

| Benchmark | Claude Haiku 4.5 | Grok 4.7 |
|---|---|---|
| GPQA Diamond | 71.2% | 92.7% |
| SimpleQA Verified | 13.2% | 56% |
| LMArena Expert | 1442 | 1422 |
| MMLU-Pro | 77.7% | — |
| Vectara Hallucination Rate | 9.8% | — |
| GPQA (HELM) | 60.5% | — |

## Multimodal

- Claude Haiku 4.5: 26.8 (#118)
- Grok 4.7: 35.5 (#87)

| Benchmark | Claude Haiku 4.5 | Grok 4.7 |
|---|---|---|
| Blueprint-Bench 2 | 0% | 32.5% |
| LMArena Vision | — | 1228 |
| Furniture Assembly | — | 20.8% |
| LMArena Document | 1420 | — |

## Multilingual

- Claude Haiku 4.5: 49.9 (#129)
- Grok 4.7: 50.8 (#116)

| Benchmark | Claude Haiku 4.5 | Grok 4.7 |
|---|---|---|
| LMArena Non-English | 1377 | 1389 |
| LMArena Chinese | 1417 | 1455 |
| LMArena French | 1408 | 1455 |
| LMArena Russian | 1381 | 1397 |
| LMArena Spanish | 1420 | 1400 |
| LMArena German | 1375 | — |
| LMArena Japanese | 1339 | — |
| LMArena Korean | 1347 | — |

## Instruction Following

- Claude Haiku 4.5: 71.4 (#149)
- Grok 4.7: 74.1 (#105)

| Benchmark | Claude Haiku 4.5 | Grok 4.7 |
|---|---|---|
| LMArena Instruction Following | 1414 | 1404 |
| IFEval | 80.1% | — |

## Long Context

- Claude Haiku 4.5: 43.6 (#92)
- Grok 4.7: 43.1 (#104)

| Benchmark | Claude Haiku 4.5 | Grok 4.7 |
|---|---|---|
| LMArena Longer Query | 1427 | 1413 |

## Writing & Preference

- Claude Haiku 4.5: 57.9 (#123)
- Grok 4.7: 70.0 (#24)

| Benchmark | Claude Haiku 4.5 | Grok 4.7 |
|---|---|---|
| LMArena Text | 1396 | 1399 |
| LMArena Creative Writing | 1372 | 1391 |
| LMArena Multi-Turn | 1409 | 1393 |
| EQ-Bench Creative Writing | — | 2007 |
| WildBench | 83.9% | — |
| EQ-Bench 4 | 1064 | — |

## FAQ

### Is Claude Haiku 4.5 better than Grok 4.7?

Grok 4.7 is the stronger model overall, scoring 53.1 to 39.5 on the Noometry Index. Claude Haiku 4.5 costs 1.5× less per token, which makes it the better buy when Grok 4.7's lead doesn't matter for your workload.

### Which is cheaper, Claude Haiku 4.5 or Grok 4.7?

Claude Haiku 4.5 is cheaper. It lists at $1 per million input tokens and $5 per million output tokens; Grok 4.7 lists at $2 and $6.

### Is Claude Haiku 4.5 or Grok 4.7 better for coding?

Grok 4.7 scores higher on coding benchmarks: 58.0 versus 44.0 in the Noometry coding category.

### Which has the bigger context window?

Grok 4.7 does, with 500K tokens against 200K.

### How many benchmarks do Claude Haiku 4.5 and Grok 4.7 share?

27 benchmarks have published results for both models. Claude Haiku 4.5 has 53 scored results on Noometry and Grok 4.7 has 39.