Anthropic
Claude Haiku 4.5 benchmark results
17 results on 8 benchmarks from 4 sources.
Also listed as: claude-haiku-4-5-20251001 · claude-haiku-4-5-20251001_32K · claude-haiku-4-5-20251001_unknown · Claude Haiku 4.5 (unknown thinking) · Claude 4.5 Haiku · Claude 4.5 Haiku (high) · Claude-Haiku-4-5-20251001 (FC) · Claude-Haiku-4-5-20251001 (Prompt)
All benchmark results
| Benchmarks | Score | Rank | Setting | Source type | Date |
|---|---|---|---|---|---|
| LMArena Text | 1414 | 142 | Overall, style control on; 146351 votes; 95% CI 1412–1416“claude-haiku-4-5-20251001” | Official leaderboard | Oct 8, 2026 |
| LMArena WebDev | 1330 | 117 | Overall; 29113 votes; 95% CI 1325–1335“claude-haiku-4-5-20251001” | Official leaderboard | Oct 8, 2026 |
| GPQA Diamond | 71.2% | 158 | Epoch AI run, accuracy | Independent run | Oct 22, 2025 |
| SWE-bench Verified | 66.6% | 32 | agent: mini-SWE-agent“claude-haiku-4-5-20251001” | Official leaderboard | Feb 17, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 66.7% | 138 | Epoch AI run, OTIS Mock AIME 2024–2025 | Independent run | Oct 22, 2025 |
| MATH Level 5 | 96.4% | 11 | Epoch AI run, accuracy | Independent run | Oct 22, 2025 |
| Terminal-Bench 2.0 | 35.5% | 112 | Terminal-Bench 2.0; agent: Goose; effort: unknown thinking | Official leaderboard via Epoch AI | Oct 15, 2025 |
| Terminal-Bench 2.0 | 29.8% | 129 | Terminal-Bench 2.0; agent: Mini-SWE-Agent; effort: unknown thinking | Official leaderboard via Epoch AI | Oct 15, 2025 |
| Terminal-Bench 2.0 | 29.8% | 130 | Terminal-Bench 2.0; agent: Mini-SWE-Agent | Official leaderboard via Epoch AI | Nov 3, 2025 |
| Terminal-Bench 2.0 | 28.3% | 134 | Terminal-Bench 2.0; agent: Terminus 2; effort: unknown thinking | Official leaderboard via Epoch AI | Oct 15, 2025 |
| Terminal-Bench 2.0 | 28.3% | 135 | Terminal-Bench 2.0; agent: Terminus 2 | Official leaderboard via Epoch AI | Oct 31, 2025 |
| Terminal-Bench 2.0 | 27.5% | 138 | Terminal-Bench 2.0; agent: Claude Code; effort: unknown thinking | Official leaderboard via Epoch AI | Oct 15, 2025 |
| Terminal-Bench 2.0 | 27.5% | 139 | Terminal-Bench 2.0; agent: Claude Code | Official leaderboard via Epoch AI | Nov 4, 2025 |
| Terminal-Bench 2.0 | 13.9% | 183 | Terminal-Bench 2.0; agent: OpenHands; effort: unknown thinking | Official leaderboard via Epoch AI | Oct 15, 2025 |
| Terminal-Bench 2.0 | 13.9% | 184 | Terminal-Bench 2.0; agent: OpenHands | Official leaderboard via Epoch AI | Nov 2, 2025 |
| BFCL v4 (function calling) | 68.7% | 6 | Overall accuracy; native function calling (FC)“Claude-Haiku-4-5-20251001 (FC)” | Official leaderboard | Oct 17, 2026 |
| BFCL v4 (function calling) | 25.3% | 87 | Overall accuracy; prompt-based“Claude-Haiku-4-5-20251001 (Prompt)” | Official leaderboard | Oct 17, 2026 |
FAQ
Where do these Claude Haiku 4.5 scores come from?
From Epoch AI Benchmarking Hub, Berkeley Function Calling Leaderboard (BFCL), LMArena leaderboard dataset, and SWE-bench leaderboards (swebench.com). Each row links to the original leaderboard, dataset or report, and vendor-reported numbers are labelled.