Across 11 comparable benchmarks, Haiku 4.5 has better reported results on 9 and worse on 2.
Moderate evidence11 of 11 shared results countedHow we compare
Price basis: Listed rates vs standard rates. The price section below shows the applicable periods.
| Fact | Qwen3-30B-A3B-2507 | Haiku 4.5 |
|---|---|---|
| Context window | 256K tokens | 200K tokens |
| Input / 1M | — | $1.00 |
| Output / 1M | — | $5.00 |
| Blended (4:1) | — | $1.15 |
| Status | Deprecated | Available |
| Released | July 29, 2025 | October 15, 2025 |
Blended is 4:1 input to output with 90% cache hits at listed rates — set your own mix.
Mean of 6 shared benchmarks grouped by capability on a 0–100 scale. How averages are built
Haiku 4.5 ahead on 3 of 3 counted
Haiku 4.5 ahead on 1 of 1 counted
Haiku 4.5 ahead on 2 of 2 counted
Haiku 4.5 ahead on 1 of 1 counted
Split 1–1 of 2 counted
No shared comparable results for Abstract reasoning, Computer & tool use, Research & search, Long context, Cybersecurity.
Published scores, with matching modes where available. Gaps show Qwen3-30B-A3B-2507 relative to Haiku 4.5. Reading the results
| Benchmark | Qwen3-30B-A3B-2507 | Haiku 4.5 | Gap |
|---|---|---|---|
| LiveCodeBenchAlgorithmic Coding · Pass @K | 43.2Normal · no tools | 62.0≠Extended · no tools | 0.70×#92 vs #69 of 113 |
| SWE-bench VerifiedRepository Engineering · Accuracy | 22.0Thinking · no tools | 73.3≠Thinking · tools · 128K | −51.3 pts#107 vs #47 of 108 |
| Terminal Bench HardAgentic Development · Accuracy | 5.3Thinking · tools | 27.3≠Normal · tools | −22.0 pts#135 vs #74 of 157 |
| AIME2025Mathematics · Accuracy | 61.3Normal · no tools | 96.3≠Thinking · tools · 128K | −35.0 pts#75 vs #18 of 86 |
| GPQA DiamondScientific Reasoning · Accuracy | 70.4Normal · no tools | 73.3≠Extended · no tools | −2.9 pts#138 vs #126 of 200 |
| SciCodeScientific Computing · Score | 33.0Thinking · no tools | 42.2Thinking · no tools | 0.78×#89 vs #72 of 98 |
| GDP.pdfDocuments & Charts · Task pass rate (%) | 0.4Thinking · no tools | 3.8Thinking · no tools | −3.4 pts#96 vs #78 of 96 |
| τ²-BenchService Workflows · Accuracy | 49.0Thinking · tools | 33.0≠Normal · tools | +16.0 pts#33 vs #39 of 40 |
| τ³-BankingService Workflows · Score | 5.4Thinking · tools | 9.3Thinking · tools | 0.58×#107 vs #95 of 120 |
| HLEKnowledge Exams · Accuracy | 9.8Thinking · no tools | 9.7≠Extended · no tools | +0.1 pts#128 vs #129 of 151 |
| MMLU-ProKnowledge Exams · Accuracy | 78.4Normal · no tools | 80.0Normal · no tools | −1.6 pts#69 vs #62 of 153 |
Algorithmic Coding · Pass @K
Repository Engineering · Accuracy
Agentic Development · Accuracy
Mathematics · Accuracy
Scientific Reasoning · Accuracy
Scientific Computing · Score
Documents & Charts · Task pass rate (%)
Service Workflows · Accuracy
Service Workflows · Score
A dash means no result is available.
| Benchmark | Qwen3-30B-A3B-2507 | Haiku 4.5 |
|---|---|---|
| CritPt | 0.3 | — |
| FrontierMath | — | 4.1 |
| FrontierMath - Tier 4 | — | 2.1 |
| Code Migration | — | 7.5 |
| SWE-Bench Pro - Public | — | 39.5 |
| SWE-Bench Pro V2 Hard | — | 25.5 |
| Terminal-Bench 1.0 | — | 41.8 |
| Vibe Code Bench v1.1 | — | 11.4 |
| Claw Bench | — | 89.4 |
| MCP-Atlas | — | 40.2 |
| OSWorld 2.1 | — | 15.7 |
| Pinch Bench | — | 82.0 |
| PinchBench v2 | — | 67.7 |
| ExploitBench | — | 13.7 |
| AA Intelligence Index v5.0.1 | — | 7.4 |
| ECI | — | 142 |
| LiveBench | — | 61.3 |
| LiveBench 2.0 | — | 46.0 |
| Vals Index | — | 22.9 |
| IF Bench | — | 54.3 |
| Context Arena | — | 36.3 |
| MMMU | — | 73.2 |
| MMMU-Pro | — | 58.6 |
| AA-AnalystAgent | — | 15.0 |
| AutomationBench | — | 0.46 |
| EMB (Excel Modeling) | — | 22.7 |
| Finance Agent v2 | — | 31.0 |
| Harvey Lab-AA | — | 61.0 |
| Harvey's Legal Agent Benchmark | — | 0.83 |
| Legal Research Bench | — | 10.6 |
| Public Benefits Bench v1.1 | — | 54.3 |
| SAGE | — | 31.8 |
| ARC-AGI-1 | — | 47.7 |
| ARC-AGI-2 | — | 4.5 |
| BALROG | — | 31.2 |
| Vending-Bench 2 | — | 459 |
| BrowseComp | 2.3 | — |
| GAIA | — | 56.4 |
| MedCode | — | 32.7 |
| MedScribe | — | 85.2 |
Time-of-day providers use peak-hour rates; fixed-price providers use standard rates. Off-peak discounts remain in the published details.
Drag the handle above the bar for cache hits, the one below it for the input : output mix · 5% steps
Blended price = effective input rate at 90% cache hits × 80% + output rate × 20%
A blended cost needs a published input and output price on both sides, and one of these two does not have them yet.
Qwen3-30B-A3B-2507: Listed text rates; pricing tier not confirmed.
Haiku 4.5: Standard pricing. Anthropic official pricing (platform.claude.com/docs/en/about-claude/pricing) verified 2026-11-13
| Price / 1M tokens | Qwen3-30B-A3B-2507 | Haiku 4.5 | Ratio |
|---|---|---|---|
| Input | — | $1.00 | — |
| Output | — | $5.00 | — |
| Cached input | — | $0.10 | — |
| Cache write | — | — | — |
| Blended | — | $1.15 | — |
Estimates use the selected token mix. Cache writes, tool fees and long-context surcharges are excluded.
Reported cost per task, paired with the score from the same run. Settings and sources may differ; costs do not affect the verdict.
| Benchmark | Qwen3-30B-A3B-2507 | Haiku 4.5 |
|---|---|---|
| AA Intelligence Index v5.0.1 | Not reported | |
| ARC-AGI-1 | Not reported | |
| AutomationBench | Not reported | $0.07 / task Average · Pass Rate: 0.46 Reported best (effort unspecified) · zapier-private-heldout |
| LiveBench 2.0 | Not reported | $0.934 / task Average · Overall score (equal-weight mean of 7 suite scores, 0-100): 46.0 Max (task-specific tools) |
| Public Benefits Bench v1.1 | Not reported | $0.2245 / task Average · Criterion pass rate (%): 54.3 Unspecified Reasoning (With Tools) |
| Attribute | Qwen3-30B-A3B-2507 | Haiku 4.5 |
|---|---|---|
| Availability | Deprecated | Available |
| Provider | Alibaba | Anthropic |
| Model type | Chat model | Multimodal model |
| First release | July 29, 2025 | October 15, 2025 |
| Context window | 256K tokens | 200K tokens |
| Max output | 16K tokens | 64K tokens |
| Total parameters | 30.5B | — |
| Active parameters | 3.3B | — |
| Architecture | Mixture of Experts (MoE) | Undisclosed |
| Documented capabilities | Multilingual | Reasoning, Multilingual |
| Input modalities | Text | Text, Image |
| Output modalities | Text | Text |
| Weights | Open weights | Closed / API only |
| Licence | Apache 2.0 | — |
| Commercial use | Commercial use allowed | Proprietary |
These are reported-score comparisons, not controlled head-to-head tests.
What people found comparing these models, or using either one. Comments also appear on the models' own pages.
Used Qwen3-30B-A3B-2507 or Haiku 4.5? Sign in with Google to share what worked, what did not, and which API or setup you tested.
No community comments on either model yet.
Alternatives with the most benchmark overlap for this matchup.