Model comparison
Claude Haiku 4.5 vs Grok 4.6
Grok 4.6 is the stronger model overall, scoring 56.9 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 4.6's lead doesn't matter for your workload.
Last verified . 35 shared benchmarks.
Summary
- They share 35 benchmarks with published results for both. Claude Haiku 4.5 scores higher in 0 categories and Grok 4.6 in 10 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Grok 4.6 leads 61.4 to 15.1.
- The biggest single-benchmark swing is NYT Connections (extended): 14.3% for Claude Haiku 4.5 and 80% for Grok 4.6.
- Claude Haiku 4.5 is cheaper at $1 / $5 per million input/output tokens, against $2 / $6 for Grok 4.6.
- Grok 4.6 accepts more context: 500K tokens versus 200K.
Side by side
| Claude Haiku 4.5 | Grok 4.6 | |
|---|---|---|
| Provider | Anthropic | xAI |
| Noometry Index | 39.5 | 56.9 |
| Released | 2025-10-15 | 2026-08-12 |
| Weights | Proprietary | Proprietary |
| Context window | 200K | 500K |
| Max output | 64K | 500K |
| Input $ / M tokens | $1 | $2 |
| Output $ / M tokens | $5 | $6 |
| Results tracked | 53 | 49 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Grok 4.6 leads
Claude Haiku 4.5: 44.0 (#78), Grok 4.6: 58.5 (#16)
| Benchmark | Claude Haiku 4.5 | Grok 4.6 |
|---|---|---|
| LMArena WebDev | 1330 | 1617 |
| SciCode | 43.3% | 56.5% |
| WeirdML | 45.4% | 67.3% |
| LMArena Coding | 1453 | 1465 |
| ALE-Bench | 653.48 | 1,508 |
| DeepSWE | — | 67.5% |
| FrontierCode | — | 48% |
| SWE-bench Verified (bash only) | 66.6% | — |
| CursorBench | — | 41.4% |
| SWE-bench Multilingual | 64.7% | — |
| FrontierSWE | — | 25.3% |
Agentic & Tool Use Grok 4.6 leads
Claude Haiku 4.5: 33.6 (#52), Grok 4.6: 39.4 (#27)
| Benchmark | Claude Haiku 4.5 | Grok 4.6 |
|---|---|---|
| Vending-Bench 2 | 458.89 | 9,047 |
| Terminal-Bench | 35.5% | — |
| APEX-Agents | — | 65.3% |
| Berkeley Function Calling Leaderboard | 68.7% | — |
| DeepResearch Bench | 45.5% | — |
| BALROG | 31.2% | — |
| ExploitBench | 13.7% | — |
| GDP.pdf | — | 17.2% |
Reasoning Grok 4.6 leads
Claude Haiku 4.5: 15.1 (#320), Grok 4.6: 61.4 (#20)
| Benchmark | Claude Haiku 4.5 | Grok 4.6 |
|---|---|---|
| ARC-AGI-2 | 4% | 67.1% |
| NYT Connections (extended) | 14.3% | 80% |
| ARC-AGI-1 | 47.7% | 87.5% |
| CritPt | 0% | 19.7% |
| Chess Puzzles | 8% | 40% |
| LMArena Hard Prompts | 1420 | 1447 |
| DTBench | 73.6% | 97.3% |
| LMCA | 30.9% | 48.5% |
| Epoch Capabilities Index | 142.41 | 156.44 |
| SimpleBench | — | 75.9% |
| EBR-Bench | — | 30.5% |
| Mystery Game Puzzles | — | 34% |
| ForecastBench | 61.4 | — |
Math Grok 4.6 leads
Claude Haiku 4.5: 44.9 (#78), Grok 4.6: 67.0 (#24)
| Benchmark | Claude Haiku 4.5 | Grok 4.6 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.7% | 99.2% |
| LMArena Math | 1396 | 1423 |
| FrontierMath (Tiers 1-3) | — | 66% |
| FrontierMath Tier 4 | — | 31.7% |
| ProofBench | — | 51% |
| Omni-MATH | 56.1% | — |
| MATH Level 5 | 96.4% | — |
| FrontierMath (Feb 2025 set) | 5.9% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge Grok 4.6 leads
Claude Haiku 4.5: 37.7 (#153), Grok 4.6: 63.3 (#20)
| Benchmark | Claude Haiku 4.5 | Grok 4.6 |
|---|---|---|
| GPQA Diamond | 71.2% | 94% |
| SimpleQA Verified | 13.2% | 49.3% |
| LMArena Expert | 1442 | 1467 |
| MMLU-Pro | 77.7% | — |
| Vectara Hallucination Rate | 9.8% | — |
| GPQA (HELM) | 60.5% | — |
Multimodal Grok 4.6 leads
Claude Haiku 4.5: 26.8 (#118), Grok 4.6: 43.6 (#23)
| Benchmark | Claude Haiku 4.5 | Grok 4.6 |
|---|---|---|
| Blueprint-Bench 2 | 0% | 33.2% |
| LMArena Document | 1420 | 1452 |
| LMArena Vision | — | 1263 |
| Furniture Assembly | — | 40% |
Multilingual Grok 4.6 leads
Claude Haiku 4.5: 49.9 (#129), Grok 4.6: 53.0 (#74)
| Benchmark | Claude Haiku 4.5 | Grok 4.6 |
|---|---|---|
| LMArena Non-English | 1377 | 1420 |
| LMArena Chinese | 1417 | 1480 |
| LMArena French | 1408 | 1461 |
| LMArena German | 1375 | 1431 |
| LMArena Japanese | 1339 | 1376 |
| LMArena Korean | 1347 | 1397 |
| LMArena Russian | 1381 | 1422 |
| LMArena Spanish | 1420 | 1404 |
Instruction Following Grok 4.6 leads
Claude Haiku 4.5: 71.4 (#149), Grok 4.6: 75.4 (#63)
| Benchmark | Claude Haiku 4.5 | Grok 4.6 |
|---|---|---|
| LMArena Instruction Following | 1414 | 1431 |
| IFEval | 80.1% | — |
Long Context Too close to call
Claude Haiku 4.5: 43.6 (#92), Grok 4.6: 44.5 (#66)
| Benchmark | Claude Haiku 4.5 | Grok 4.6 |
|---|---|---|
| LMArena Longer Query | 1427 | 1454 |
Writing & Preference Grok 4.6 leads
Claude Haiku 4.5: 57.9 (#123), Grok 4.6: 62.3 (#80)
| Benchmark | Claude Haiku 4.5 | Grok 4.6 |
|---|---|---|
| LMArena Text | 1396 | 1428 |
| LMArena Creative Writing | 1372 | 1428 |
| LMArena Multi-Turn | 1409 | 1425 |
| WildBench | 83.9% | — |
| EQ-Bench 4 | 1064 | — |
Frequently asked questions
Is Claude Haiku 4.5 better than Grok 4.6?
Grok 4.6 is the stronger model overall, scoring 56.9 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 4.6's lead doesn't matter for your workload.
Which is cheaper, Claude Haiku 4.5 or Grok 4.6?
Claude Haiku 4.5 is cheaper. It lists at $1 per million input tokens and $5 per million output tokens; Grok 4.6 lists at $2 and $6.
Is Claude Haiku 4.5 or Grok 4.6 better for coding?
Grok 4.6 scores higher on coding benchmarks: 58.5 versus 44.0 in the Noometry coding category.
Which has the bigger context window?
Grok 4.6 does, with 500K tokens against 200K.
How many benchmarks do Claude Haiku 4.5 and Grok 4.6 share?
35 benchmarks have published results for both models. Claude Haiku 4.5 has 53 scored results on Noometry and Grok 4.6 has 49.