Agentic work
Tool use, computer use, and multi-step task completion
Qwen3.7 Max
Qwen3.7 Max has the higher public agentic point estimate, 40.1 to 34, with Supported evidence for both models. The conditional ranges do not establish rank confidence.
Updated November 17, 2026. Rank says Qwen3.7 Max is ahead. Price, access, and your workload can each overturn that. Public scores include evidence status and uncertainty.
Qwen3.7 Max has the higher public point estimate, 60.78 versus 56.58. Their conditional score ranges overlap. These ranges do not establish rank confidence. 12 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
Moonshot AI
56.58/100
Estimated · Public rank #69
Conditional range 42.2–70.9
Kimi K2.5 (Reasoning): $0.60 input / $3 output
Qwen3.7 Max: No comparable API price
Kimi K2.5 (Reasoning): 256K
Qwen3.7 Max: 1M
Kimi K2.5 (Reasoning): January 27, 2026
Qwen3.7 Max: May 16, 2026
Recommendations appear only when a shared evidence basis or an explicit operating constraint supports the call. Secondary and unsupported use cases stay disclosed below the initial list.
Tool use, computer use, and multi-step task completion
Qwen3.7 Max
Qwen3.7 Max has the higher public agentic point estimate, 40.1 to 34, with Supported evidence for both models. The conditional ranges do not establish rank confidence.
Prompts that approach the documented context limit
Qwen3.7 Max
Qwen3.7 Max has the larger documented context window.
Code generation, repair, and software-engineering tasks
Not enough matched evidence
Kimi K2.5 (Reasoning) is scored on Estimated evidence for coding, so the reading is directional rather than like-for-like.
1K fresh input + 500 output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
200K cached + 20K fresh input + 10K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
50K fresh input + 3K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Choose a job from the LLM Selector's task catalog to see the category row it rests on, under the same basis rules as the table below. A directional row stays directional; choosing a job never creates a winner.
The same task catalog as the LLM Selector. Each job names the evidence surface it rests on; nothing here adds a new score.
Directional only · BenchAlign v5.8
Qwen3.7 Max has the higher coding point estimate. Conditional score ranges do not establish rank confidence.
Coding scores combine specific tasks and setups. Match the editor, harness, and effort to your workflow.
The LLM Selector reads what you need a model for and returns a shortlist, ranked in the same order the job pages publish.
We keep what you type here, drafts included, for 180 days, and after that only the job and requirements read from it. Nothing is tied to your IP address or account. Leave out names and other personal details. Privacy
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
1 category rests on Estimated evidence or different evidence sets. Those rows remain visible for coverage context but do not name a winner.
Only shared public evidence is shown. Sparse evidence stays a ruled list rather than being closed into a radar shape.
Too few matched category axes support a radar. The ruled list below shows only shared benchmark results; positions use each benchmark’s normalized display scale when available.
Terminal-Bench 2.0Agentic
Normalized gap 18.9GPQAKnowledge
Normalized gap 4.8SWE-bench VerifiedCoding
Normalized gap 3.6MMLU-Pro (Vals)Knowledge
Normalized gap 3.4LiveCodeBench (Vals)Coding
Normalized gap 3.2Each row shows the public-lane category score for both models: the BenchAlign v5.8 lane for agentic, coding, reasoning, multimodal, and knowledge, and the weighted public rows elsewhere, with the same rank each model page reports. A row is like-for-like only when both scores rest on Supported evidence or the same weighted set. Directional and not-comparable rows remain visible, but they never receive a winner.
| Category | Kimi K2.5 (Reasoning) | Qwen3.7 Max | Basis | Reading |
|---|---|---|---|---|
| Agentic | 34.0Supported · #93/138 | 40.1Supported · #75/138 | Like-for-likeBenchAlign v5.8 lane · 5 vs 10 public rows | Qwen3.7 Max leads · intervals overlap |
| Knowledge | 52.2Supported · #69/181 | 60.2Supported · #48/181 | Like-for-likeBenchAlign v5.8 lane · 4 vs 9 public rows | Qwen3.7 Max leads · intervals overlap |
| Coding | 34.8Estimated · #95/163 | 44.8Supported · #62/163 | Directional onlyBenchAlign v5.8 lane · 6 vs 10 public rows | Directional only |
| Reasoning | Not ranked | 51.7Supported · #48/160 | Not comparableBenchAlign v5.8 lane · 1 vs 2 public rows | Not comparable |
| Multimodal | 53.4Estimated · #47/67 | Not ranked | Not comparableBenchAlign v5.8 lane · 2 vs 0 public rows | Not comparable |
| Multilingual | Not ranked | 100.0#1/17 | Not comparableProvisional lane · 0 vs 1 weighted rows | Not comparable |
| Instruction following | Not ranked | 92.8#17/157 | Not comparableProvisional lane · 0 vs 1 weighted rows | Not comparable |
| Math | Not ranked | 78.3#2/12 | Not comparableProvisional lane · 0 vs 3 weighted rows | Not comparable |
Ranks count the models scored in each category’s lane, so the agentic, coding, reasoning, multimodal, and knowledge denominators (BenchAlign v5.8) differ from the provisional-lane categories. Unranked scores sit on the provisional lane’s scale but never name a winner. Like-for-like rows within 0.5 points read as a practical tie.
Three fixed token mixes turn per-token rates into comparable decisions. Each scenario states context fit and whether cached input had to fall back to the published list-input rate.
1K fresh input + 500 output tokens
Qwen3.7 Max has no comparable published API token rate.
50K fresh input + 3K output tokens
Qwen3.7 Max has no comparable published API token rate.
200K cached + 20K fresh input + 10K output tokens
Kimi K2.5 (Reasoning) has no published cached-input rate, so cached tokens use its listed input rate. Qwen3.7 Max has no comparable published API token rate.
Overall score against output price per 1M tokens, one model per family. Pricing as of November 17, 2026.
Kimi K2.5 (Reasoning)
Gray dots are the other priced models. The dark step line is the efficiency frontier: for a model under it, another model costs no more and scores at least as high.
See Kimi K2.5 (Reasoning) on the full price vs performance chartSourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
Maximum documented context; output-token limits may be lower.
Kimi K2.5 (Reasoning)
256K
Qwen3.7 Max
1M
Kimi K2.5 (Reasoning)
Not sourced
Qwen3.7 Max
Not sourced
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
Kimi K2.5 (Reasoning)
Not published
Qwen3.7 Max
No comparable hosted API rate
Kimi K2.5 (Reasoning)
text, image, video
Kimi-K2.5 model cardQwen3.7 Max
Kimi K2.5 (Reasoning)
Qwen3.7 Max
Kimi K2.5 (Reasoning)
Not sourced
Qwen3.7 Max
Not sourced
Kimi K2.5 (Reasoning)
Reasoning
Qwen3.7 Max
Reasoning
Kimi K2.5 (Reasoning)
Open Weight
Qwen3.7 Max
Proprietary
Kimi K2.5 (Reasoning)
Open Weight
Qwen3.7 Max
Proprietary
Kimi K2.5 (Reasoning)
2026-01-27
Qwen3.7 Max
2026-05-16
Qwen3.7 Max has the higher public point estimate, 60.78 versus 56.58. Their conditional score ranges overlap. These ranges do not establish rank confidence. The page therefore keeps the decision tied to the specific documented workload.
Qwen3.7 Max scores higher for coding on the public lane, 44.8 to 34.8. Kimi K2.5 (Reasoning) is scored on Estimated evidence for coding, so the reading is directional rather than like-for-like. BenchLM does not name a winner for coding from a directional reading; read the shared benchmark rows directly and test the models on the same task set.
Qwen3.7 Max has the higher public agentic tasks point estimate, 40.1 to 34, with Supported evidence for both models. The conditional ranges do not establish rank confidence.
Both models do not have comparable published API token rates, so this page does not name a universal price winner.
Qwen3.7 Max has the larger documented context window: 1M, compared with 256K.
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Terminal-Bench 2.0
Qwen3.7 Max leads this result
BrowseComp
Not directly comparable
OSWorld-Verified
Not directly comparable
Gert Labs
Shared sourceQwen3.7 Max leads this result
Terminal-Bench 2.1 (Vals)
Qwen3.7 Max leads this result
QwenClawBench
Not directly comparable
Claw-Eval
Not directly comparable
BFCL v4
Not directly comparable
MCP Atlas
Not directly comparable
VITA-Bench
Not directly comparable
HLE w/ tools
Not directly comparable
ResearchClawBench
Not directly comparable
SWE-bench Verified
Qwen3.7 Max leads this result
LiveCodeBench v6
Qwen3.7 Max leads this result
Terminal-Bench 2.0
Qwen3.7 Max leads this result
Vibe Code Bench
Not directly comparable
LiveCodeBench (Vals)
Qwen3.7 Max leads this result
SWE-bench (Vals)
Kimi K2.5 (Reasoning) leads this result
SWE-bench Pro
Not directly comparable
SWE Multilingual
Not directly comparable
NL2Repo
Not directly comparable
SciCode
Not directly comparable
OpenHarmony Bench v1.0
Not directly comparable
GPQA
Qwen3.7 Max leads this result
MMLU-Pro
Qwen3.7 Max leads this result
GPQA Diamond (Vals)
Qwen3.7 Max leads this result
MMLU-Pro (Vals)
Qwen3.7 Max leads this result
GPQA-D
Not directly comparable
HLE
Not directly comparable
MMLU-Redux
Not directly comparable
SuperGPQA
Not directly comparable
MMMLU
Not directly comparable
MMLU-ProX
Not directly comparable
NOVA-63
Not directly comparable
INCLUDE
Not directly comparable
MAXIFE
Not directly comparable
PolyMath
Not directly comparable
AIME 2025
Not directly comparable
HMMT Feb 2025
Not directly comparable
HMMT Feb 2026
Not directly comparable
IMOAnswerBench
Not directly comparable
Apex
Not directly comparable
FrontierMath v2 (Tiers 1-3)
Not directly comparable
FrontierMath v2 (Tier 4)
Not directly comparable
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 7,000+ subscribers.
Last updated November 17, 2026