Model comparison
Claude Haiku 4.5 vs DeepSeek-V2.5 (Sep 2024)
Claude Haiku 4.5 is the stronger model overall, scoring 39.3 to 37.0 on the Noometry Index.
Last verified . 17 shared benchmarks.
DeepSeekSummary
- They share 17 benchmarks with published results for both. Claude Haiku 4.5 scores higher in 7 categories and DeepSeek-V2.5 (Sep 2024) in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in coding, where Claude Haiku 4.5 leads 43.9 to 31.3.
- DeepSeek-V2.5 (Sep 2024) has downloadable open weights; the other is API-only.
Side by side
| Claude Haiku 4.5 | DeepSeek-V2.5 (Sep 2024) | |
|---|---|---|
| Provider | Anthropic | DeepSeek |
| Noometry Index | 39.3 | 37.0 |
| Released | 2025-10-15 | 2024-09-06 |
| Weights | Proprietary | Open |
| Context window | 200K | — |
| Max output | 64K | — |
| Input $ / M tokens | $1 | — |
| Output $ / M tokens | $5 | — |
| Results tracked | 53 | 22 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Haiku 4.5 leads
Claude Haiku 4.5: 43.9 (#71), DeepSeek-V2.5 (Sep 2024): 31.3 (#274)
| Benchmark | Claude Haiku 4.5 | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Coding | 1454 | 1310 |
| SWE-bench Verified (bash only) | 66.6% | — |
| Aider Polyglot | — | 17.8% |
| LMArena WebDev | 1330 | — |
| SWE-bench Multilingual | 64.7% | — |
| SciCode | 43.3% | — |
| WeirdML | 45.4% | — |
| BigCodeBench Instruct | — | 48.6% |
| BigCodeBench Complete | — | 53.2% |
| ALE-Bench | 653.48 | — |
| HumanEval+ | — | 83.5% |
| MBPP+ | — | 74.1% |
Agentic & Tool Use Not comparable
Claude Haiku 4.5: 33.5 (#50), DeepSeek-V2.5 (Sep 2024): —
| Benchmark | Claude Haiku 4.5 | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| Terminal-Bench | 35.5% | — |
| Berkeley Function Calling Leaderboard | 68.7% | — |
| DeepResearch Bench | 45.5% | — |
| BALROG | 31.2% | — |
| ExploitBench | 13.7% | — |
| Vending-Bench 2 | 458.89 | — |
Reasoning DeepSeek-V2.5 (Sep 2024) leads
Claude Haiku 4.5: 15.1 (#310), DeepSeek-V2.5 (Sep 2024): 24.5 (#145)
| Benchmark | Claude Haiku 4.5 | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Hard Prompts | 1420 | 1289 |
| ARC-AGI-2 | 4% | — |
| NYT Connections (extended) | 14.3% | — |
| ARC-AGI-1 | 47.7% | — |
| CritPt | 0% | — |
| Chess Puzzles | 8% | — |
| DTBench | 73.6% | — |
| LMCA | 30.9% | — |
| Epoch Capabilities Index | 142.42 | — |
| ForecastBench | 61.4 | — |
Math Claude Haiku 4.5 leads
Claude Haiku 4.5: 43.0 (#69), DeepSeek-V2.5 (Sep 2024): 33.3 (#177)
| Benchmark | Claude Haiku 4.5 | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Math | 1395 | 1289 |
| OTIS Mock AIME 2024-2025 | 66.7% | — |
| 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 leads
Claude Haiku 4.5: 37.6 (#146), DeepSeek-V2.5 (Sep 2024): 34.2 (#190)
| Benchmark | Claude Haiku 4.5 | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Expert | 1443 | 1266 |
| GPQA Diamond | 71.2% | — |
| SimpleQA Verified | 13.2% | — |
| MMLU-Pro | 77.7% | — |
| Vectara Hallucination Rate | 9.8% | — |
| GPQA (HELM) | 60.5% | — |
Multimodal Not comparable
Claude Haiku 4.5: 27.0 (#111), DeepSeek-V2.5 (Sep 2024): —
| Benchmark | Claude Haiku 4.5 | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| Blueprint-Bench 2 | 0% | — |
| LMArena Document | 1420 | — |
Multilingual Claude Haiku 4.5 leads
Claude Haiku 4.5: 50.1 (#124), DeepSeek-V2.5 (Sep 2024): 42.7 (#188)
| Benchmark | Claude Haiku 4.5 | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Non-English | 1376 | 1273 |
| LMArena Chinese | 1415 | 1317 |
| LMArena French | 1408 | 1288 |
| LMArena German | 1377 | 1258 |
| LMArena Japanese | 1339 | 1228 |
| LMArena Korean | 1346 | 1209 |
| LMArena Russian | 1380 | 1289 |
| LMArena Spanish | 1420 | 1248 |
Instruction Following Claude Haiku 4.5 leads
Claude Haiku 4.5: 71.6 (#144), DeepSeek-V2.5 (Sep 2024): 67.7 (#189)
| Benchmark | Claude Haiku 4.5 | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Instruction Following | 1414 | 1280 |
| IFEval | 80.1% | — |
Long Context Claude Haiku 4.5 leads
Claude Haiku 4.5: 43.6 (#86), DeepSeek-V2.5 (Sep 2024): 39.5 (#169)
| Benchmark | Claude Haiku 4.5 | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Longer Query | 1427 | 1301 |
Writing & Preference Claude Haiku 4.5 leads
Claude Haiku 4.5: 58.1 (#119), DeepSeek-V2.5 (Sep 2024): 50.0 (#182)
| Benchmark | Claude Haiku 4.5 | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Text | 1396 | 1294 |
| LMArena Creative Writing | 1371 | 1285 |
| LMArena Multi-Turn | 1409 | 1297 |
| WildBench | 83.9% | — |
| EQ-Bench 4 | 1064 | — |
Frequently asked questions
Is Claude Haiku 4.5 better than DeepSeek-V2.5 (Sep 2024)?
Claude Haiku 4.5 is the stronger model overall, scoring 39.3 to 37.0 on the Noometry Index.
Is Claude Haiku 4.5 or DeepSeek-V2.5 (Sep 2024) better for coding?
Claude Haiku 4.5 scores higher on coding benchmarks: 43.9 versus 31.3 in the Noometry coding category.
How many benchmarks do Claude Haiku 4.5 and DeepSeek-V2.5 (Sep 2024) share?
17 benchmarks have published results for both models. Claude Haiku 4.5 has 53 scored results on Noometry and DeepSeek-V2.5 (Sep 2024) has 22.