Model comparison
Claude Haiku 4.5 vs MiniMax-M2.5
Claude Haiku 4.5 is the stronger model overall, scoring 39.5 to 38.3 on the Noometry Index. MiniMax-M2.5 costs 3.8× less per token, which makes it the better buy when Claude Haiku 4.5's lead doesn't matter for your workload.
Last verified . 28 shared benchmarks.
MiniMaxSummary
- They share 28 benchmarks with published results for both. Claude Haiku 4.5 scores higher in 5 categories and MiniMax-M2.5 in 4 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Claude Haiku 4.5 leads 44.9 to 26.9.
- The biggest single-benchmark swing is ARC-AGI-1: 47.7% for Claude Haiku 4.5 and 63.7% for MiniMax-M2.5.
- MiniMax-M2.5 is cheaper at $0.30 / $1.20 per million input/output tokens, against $1 / $5 for Claude Haiku 4.5.
- MiniMax-M2.5 accepts more context: 205K tokens versus 200K.
- MiniMax-M2.5 has downloadable open weights; the other is API-only.
Side by side
| Claude Haiku 4.5 | MiniMax-M2.5 | |
|---|---|---|
| Provider | Anthropic | MiniMax |
| Noometry Index | 39.5 | 38.3 |
| Released | 2025-10-15 | 2026-02-12 |
| Weights | Proprietary | Open |
| Context window | 200K | 205K |
| Max output | 64K | 131K |
| Input $ / M tokens | $1 | $0.30 |
| Output $ / M tokens | $5 | $1.20 |
| Results tracked | 53 | 33 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiniMax-M2.5 leads
Claude Haiku 4.5: 44.0 (#78), MiniMax-M2.5: 48.1 (#58)
| Benchmark | Claude Haiku 4.5 | MiniMax-M2.5 |
|---|---|---|
| SWE-bench Verified (bash only) | 66.6% | 75.8% |
| LMArena WebDev | 1330 | 1387 |
| SWE-bench Multilingual | 64.7% | 68.3% |
| LMArena Coding | 1453 | 1381 |
| ALE-Bench | 653.48 | 618.17 |
| SciCode | 43.3% | — |
| WeirdML | 45.4% | — |
Agentic & Tool Use Claude Haiku 4.5 leads
Claude Haiku 4.5: 33.6 (#52), MiniMax-M2.5: 30.4 (#77)
| Benchmark | Claude Haiku 4.5 | MiniMax-M2.5 |
|---|---|---|
| Terminal-Bench | 35.5% | 42.7% |
| Vending-Bench 2 | 458.89 | -23.16 |
| Berkeley Function Calling Leaderboard | 68.7% | — |
| DeepResearch Bench | 45.5% | — |
| BALROG | 31.2% | — |
| ExploitBench | 13.7% | — |
Reasoning MiniMax-M2.5 leads
Claude Haiku 4.5: 15.1 (#320), MiniMax-M2.5: 17.5 (#292)
| Benchmark | Claude Haiku 4.5 | MiniMax-M2.5 |
|---|---|---|
| ARC-AGI-2 | 4% | 4.9% |
| NYT Connections (extended) | 14.3% | 16.8% |
| ARC-AGI-1 | 47.7% | 63.7% |
| LMArena Hard Prompts | 1420 | 1372 |
| Epoch Capabilities Index | 142.41 | 146.68 |
| Kagi LLM Benchmark | — | 55.2% |
| CritPt | 0% | — |
| Chess Puzzles | 8% | — |
| DTBench | 73.6% | — |
| LMCA | 30.9% | — |
| ForecastBench | 61.4 | — |
Math Claude Haiku 4.5 leads
Claude Haiku 4.5: 44.9 (#78), MiniMax-M2.5: 26.9 (#253)
| Benchmark | Claude Haiku 4.5 | MiniMax-M2.5 |
|---|---|---|
| LMArena Math | 1396 | 1378 |
| OTIS Mock AIME 2024-2025 | 66.7% | — |
| ProofBench | — | 4% |
| Omni-MATH | 56.1% | — |
| MATH Level 5 | 96.4% | — |
| FrontierMath (Feb 2025 set) | 5.9% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge MiniMax-M2.5 leads
Claude Haiku 4.5: 37.7 (#153), MiniMax-M2.5: 39.2 (#135)
| Benchmark | Claude Haiku 4.5 | MiniMax-M2.5 |
|---|---|---|
| Vectara Hallucination Rate | 9.8% | 9.1% |
| LMArena Expert | 1442 | 1379 |
| GPQA Diamond | 71.2% | — |
| SimpleQA Verified | 13.2% | — |
| MMLU-Pro | 77.7% | — |
| GPQA (HELM) | 60.5% | — |
Multimodal Not comparable
Claude Haiku 4.5: 26.8 (#118), MiniMax-M2.5: —
| Benchmark | Claude Haiku 4.5 | MiniMax-M2.5 |
|---|---|---|
| Blueprint-Bench 2 | 0% | — |
| LMArena Document | 1420 | — |
Multilingual Claude Haiku 4.5 leads
Claude Haiku 4.5: 49.9 (#129), MiniMax-M2.5: 47.1 (#152)
| Benchmark | Claude Haiku 4.5 | MiniMax-M2.5 |
|---|---|---|
| LMArena Non-English | 1377 | 1338 |
| LMArena Chinese | 1417 | 1393 |
| LMArena French | 1408 | 1362 |
| LMArena German | 1375 | 1362 |
| LMArena Japanese | 1339 | 1171 |
| LMArena Korean | 1347 | 1232 |
| LMArena Russian | 1381 | 1358 |
| LMArena Spanish | 1420 | 1354 |
Instruction Following Too close to call
Claude Haiku 4.5: 71.4 (#149), MiniMax-M2.5: 71.5 (#148)
| Benchmark | Claude Haiku 4.5 | MiniMax-M2.5 |
|---|---|---|
| LMArena Instruction Following | 1414 | 1353 |
| IFEval | 80.1% | — |
Long Context Claude Haiku 4.5 leads
Claude Haiku 4.5: 43.6 (#92), MiniMax-M2.5: 37.5 (#216)
| Benchmark | Claude Haiku 4.5 | MiniMax-M2.5 |
|---|---|---|
| LMArena Longer Query | 1427 | 1366 |
| CL-bench | — | 11.4% |
| CL-bench Life | — | 6.3% |
Writing & Preference Claude Haiku 4.5 leads
Claude Haiku 4.5: 57.9 (#123), MiniMax-M2.5: 53.9 (#153)
| Benchmark | Claude Haiku 4.5 | MiniMax-M2.5 |
|---|---|---|
| LMArena Text | 1396 | 1359 |
| LMArena Creative Writing | 1372 | 1331 |
| LMArena Multi-Turn | 1409 | 1364 |
| EQ-Bench Creative Writing | — | 1361 |
| WildBench | 83.9% | — |
| EQ-Bench 4 | 1064 | — |
Frequently asked questions
Is Claude Haiku 4.5 better than MiniMax-M2.5?
Claude Haiku 4.5 is the stronger model overall, scoring 39.5 to 38.3 on the Noometry Index. MiniMax-M2.5 costs 3.8× less per token, which makes it the better buy when Claude Haiku 4.5's lead doesn't matter for your workload.
Which is cheaper, Claude Haiku 4.5 or MiniMax-M2.5?
MiniMax-M2.5 is cheaper. It lists at $0.30 per million input tokens and $1.20 per million output tokens; Claude Haiku 4.5 lists at $1 and $5.
Is Claude Haiku 4.5 or MiniMax-M2.5 better for coding?
MiniMax-M2.5 scores higher on coding benchmarks: 48.1 versus 44.0 in the Noometry coding category.
Which has the bigger context window?
MiniMax-M2.5 does, with 205K tokens against 200K.
How many benchmarks do Claude Haiku 4.5 and MiniMax-M2.5 share?
28 benchmarks have published results for both models. Claude Haiku 4.5 has 53 scored results on Noometry and MiniMax-M2.5 has 33.