Model comparison
Claude Haiku 4.5 vs DeepSeek V4.1 Flash
DeepSeek V4.1 Flash is the stronger model overall, scoring 50.3 to 39.3 on the Noometry Index.
Last verified . 22 shared benchmarks.
DeepSeekSummary
- They share 22 benchmarks with published results for both. Claude Haiku 4.5 scores higher in 0 categories and DeepSeek V4.1 Flash in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4.1 Flash leads 46.2 to 15.1.
- The biggest single-benchmark swing is NYT Connections (extended): 14.3% for Claude Haiku 4.5 and 89.6% for DeepSeek V4.1 Flash.
- DeepSeek V4.1 Flash is cheaper at $0.15 / $0.60 per million input/output tokens, against $1 / $5 for Claude Haiku 4.5.
- DeepSeek V4.1 Flash accepts more context: 1M tokens versus 200K.
- DeepSeek V4.1 Flash has downloadable open weights; the other is API-only.
Side by side
| Claude Haiku 4.5 | DeepSeek V4.1 Flash | |
|---|---|---|
| Provider | Anthropic | DeepSeek |
| Noometry Index | 39.3 | 50.3 |
| Released | 2025-10-15 | 2026-09-09 |
| Weights | Proprietary | Open |
| Context window | 200K | 1M |
| Max output | 64K | 393K |
| Input $ / M tokens | $1 | $0.15 |
| Output $ / M tokens | $5 | $0.60 |
| Results tracked | 53 | 25 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek V4.1 Flash leads
Claude Haiku 4.5: 43.9 (#71), DeepSeek V4.1 Flash: 52.8 (#29)
| Benchmark | Claude Haiku 4.5 | DeepSeek V4.1 Flash |
|---|---|---|
| LMArena WebDev | 1330 | 1620 |
| SciCode | 43.3% | 51.9% |
| LMArena Coding | 1454 | 1511 |
| ALE-Bench | 653.48 | 1,092 |
| SWE-bench Verified (bash only) | 66.6% | — |
| SWE-bench Multilingual | 64.7% | — |
| WeirdML | 45.4% | — |
Agentic & Tool Use Not comparable
Claude Haiku 4.5: 33.5 (#50), DeepSeek V4.1 Flash: —
| Benchmark | Claude Haiku 4.5 | DeepSeek V4.1 Flash |
|---|---|---|
| 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 V4.1 Flash leads
Claude Haiku 4.5: 15.1 (#310), DeepSeek V4.1 Flash: 46.2 (#39)
| Benchmark | Claude Haiku 4.5 | DeepSeek V4.1 Flash |
|---|---|---|
| NYT Connections (extended) | 14.3% | 89.6% |
| CritPt | 0% | 14.3% |
| LMArena Hard Prompts | 1420 | 1485 |
| DTBench | 73.6% | 89.9% |
| LMCA | 30.9% | 47% |
| Epoch Capabilities Index | 142.42 | 155.01 |
| ARC-AGI-2 | 4% | — |
| ARC-AGI-1 | 47.7% | — |
| Chess Puzzles | 8% | — |
| Surface Evolver Bench | — | 46.3% |
| ForecastBench | 61.4 | — |
Math DeepSeek V4.1 Flash leads
Claude Haiku 4.5: 43.0 (#69), DeepSeek V4.1 Flash: 48.4 (#52)
| Benchmark | Claude Haiku 4.5 | DeepSeek V4.1 Flash |
|---|---|---|
| LMArena Math | 1395 | 1493 |
| OTIS Mock AIME 2024-2025 | 66.7% | — |
| ProofBench | — | 54% |
| Omni-MATH | 56.1% | — |
| MATH Level 5 | 96.4% | — |
| FrontierMath (Feb 2025 set) | 5.9% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge DeepSeek V4.1 Flash leads
Claude Haiku 4.5: 37.6 (#146), DeepSeek V4.1 Flash: 41.6 (#97)
| Benchmark | Claude Haiku 4.5 | DeepSeek V4.1 Flash |
|---|---|---|
| LMArena Expert | 1443 | 1505 |
| 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 V4.1 Flash: —
| Benchmark | Claude Haiku 4.5 | DeepSeek V4.1 Flash |
|---|---|---|
| Blueprint-Bench 2 | 0% | — |
| LMArena Document | 1420 | — |
Multilingual DeepSeek V4.1 Flash leads
Claude Haiku 4.5: 50.1 (#124), DeepSeek V4.1 Flash: 55.5 (#29)
| Benchmark | Claude Haiku 4.5 | DeepSeek V4.1 Flash |
|---|---|---|
| LMArena Non-English | 1376 | 1452 |
| LMArena Chinese | 1415 | 1506 |
| LMArena French | 1408 | 1448 |
| LMArena Russian | 1380 | 1473 |
| LMArena Spanish | 1420 | 1499 |
| LMArena German | 1377 | — |
| LMArena Japanese | 1339 | — |
| LMArena Korean | 1346 | — |
Instruction Following DeepSeek V4.1 Flash leads
Claude Haiku 4.5: 71.6 (#144), DeepSeek V4.1 Flash: 77.6 (#24)
| Benchmark | Claude Haiku 4.5 | DeepSeek V4.1 Flash |
|---|---|---|
| LMArena Instruction Following | 1414 | 1475 |
| IFEval | 80.1% | — |
Long Context DeepSeek V4.1 Flash leads
Claude Haiku 4.5: 43.6 (#86), DeepSeek V4.1 Flash: 45.2 (#46)
| Benchmark | Claude Haiku 4.5 | DeepSeek V4.1 Flash |
|---|---|---|
| LMArena Longer Query | 1427 | 1476 |
Writing & Preference DeepSeek V4.1 Flash leads
Claude Haiku 4.5: 58.1 (#119), DeepSeek V4.1 Flash: 65.8 (#43)
| Benchmark | Claude Haiku 4.5 | DeepSeek V4.1 Flash |
|---|---|---|
| LMArena Text | 1396 | 1465 |
| LMArena Creative Writing | 1371 | 1436 |
| LMArena Multi-Turn | 1409 | 1460 |
| EQ-Bench Creative Writing | — | 1540 |
| WildBench | 83.9% | — |
| EQ-Bench 4 | 1064 | — |
Frequently asked questions
Is Claude Haiku 4.5 better than DeepSeek V4.1 Flash?
DeepSeek V4.1 Flash is the stronger model overall, scoring 50.3 to 39.3 on the Noometry Index.
Which is cheaper, Claude Haiku 4.5 or DeepSeek V4.1 Flash?
DeepSeek V4.1 Flash is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; Claude Haiku 4.5 lists at $1 and $5.
Is Claude Haiku 4.5 or DeepSeek V4.1 Flash better for coding?
DeepSeek V4.1 Flash scores higher on coding benchmarks: 52.8 versus 43.9 in the Noometry coding category.
Which has the bigger context window?
DeepSeek V4.1 Flash does, with 1M tokens against 200K.
How many benchmarks do Claude Haiku 4.5 and DeepSeek V4.1 Flash share?
22 benchmarks have published results for both models. Claude Haiku 4.5 has 53 scored results on Noometry and DeepSeek V4.1 Flash has 25.