Agentic
Directional only- MiniMax M3
- 42.0
- Supported · #109/152
- ZAYA1-8B
- 46.4
- Estimated · #80/152
- Basis
- BenchAlign lane · 9 vs 1 public rows
- Reading
- Directional only
MiniMax M3 vs ZAYA1-8B
Updated September 10, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.
Decision reading
The public evidence has no benchmark result shared by both models, so it does not support a quality verdict. Use the documented cost, context, and runtime rows instead.
0 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
Share or export
Which one for your work
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.
Long documents
Prompts that approach the documented context limit
MiniMax M3
MiniMax M3 has the larger documented context window.
Confidence: documented
Show secondary and unsupported calls
Coding work
Code generation, repair, and software-engineering tasks
Not enough matched evidence
ZAYA1-8B is not ranked on the public lane for coding, so no winner is named for coding.
Confidence: limited
Agentic work
Tool use, computer use, and multi-step task completion
Not enough matched evidence
ZAYA1-8B is scored on Estimated evidence for agentic, so the reading is directional rather than like-for-like.
Confidence: limited
Chat turn cost
1K fresh input + 500 output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-rates
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Not enough matched evidence
The page does not recommend a cost winner because at least one model cannot fit the stated workload in one request. ZAYA1-8B does not fit this workload in one request. ZAYA1-8B has no comparable published API token rate.
Confidence: listed-rates
Repository review cost
50K fresh input + 3K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-rates
What is actually comparable
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
27 MiniMax M3 only
11 ZAYA1-8B only
- Shared results
- 0
- MiniMax M3 only
- 27
- ZAYA1-8B only
- 11
- Like-for-like categories
- 0 / 8
3 categories rest on Estimated evidence or different evidence sets. Those rows remain visible for coverage context but do not name a winner.
Category results, on a stated basis
Each row shows the public-lane category score for both models: the BenchAlign lane for agentic, coding, 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.
Knowledge
Directional only- MiniMax M3
- 53.4
- Supported · #64/182
- ZAYA1-8B
- 43.9
- Estimated · #118/182
- Basis
- BenchAlign lane · 2 vs 3 public rows
- Reading
- Directional only
Instruction following
Directional only- MiniMax M3
- 93.7
- #4/121
- ZAYA1-8B
- 35.6
- #109/121
- Basis
- Provisional lane · 0 vs 1 weighted rows
- Reading
- Directional only
Coding
Not comparable- MiniMax M3
- 48.9
- Estimated · #66/151
- ZAYA1-8B
- Not ranked
- Basis
- BenchAlign lane · 10 vs 1 public rows
- Reading
- Not comparable
Reasoning
Not comparable- MiniMax M3
- 78.0
- Unranked · 2 rankable rows
- ZAYA1-8B
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Math
Not comparable- MiniMax M3
- Not ranked
- ZAYA1-8B
- 62.6
- Unranked · 4 rankable rows
- Basis
- Provisional lane · 1 vs 2 weighted rows
- Reading
- Not comparable
Multilingual
Not comparable- MiniMax M3
- Not ranked
- ZAYA1-8B
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Multimodal
Not comparable- MiniMax M3
- 52.2
- #34/48
- ZAYA1-8B
- Not ranked
- Basis
- Provisional lane · 2 vs 0 weighted rows
- Reading
- Not comparable
| Category | MiniMax M3 | ZAYA1-8B | Basis | Reading |
|---|---|---|---|---|
| Agentic | 42.0Supported · #109/152 | 46.4Estimated · #80/152 | Directional onlyBenchAlign lane · 9 vs 1 public rows | Directional only |
| Knowledge | 53.4Supported · #64/182 | 43.9Estimated · #118/182 | Directional onlyBenchAlign lane · 2 vs 3 public rows | Directional only |
| Instruction following | 93.7#4/121 | 35.6#109/121 | Directional onlyProvisional lane · 0 vs 1 weighted rows | Directional only |
| Coding | 48.9Estimated · #66/151 | Not ranked | Not comparableBenchAlign lane · 10 vs 1 public rows | Not comparable |
| Reasoning | 78.0Unranked · 2 rankable rows | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Math | Not ranked | 62.6Unranked · 4 rankable rows | Not comparableProvisional lane · 1 vs 2 weighted rows | Not comparable |
| Multilingual | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Multimodal | 52.2#34/48 | Not ranked | Not comparableProvisional lane · 2 vs 0 weighted rows | Not comparable |
Ranks count the models scored in each category’s lane, so the agentic, coding, and knowledge denominators (BenchAlign) differ from the provisional-lane categories. Unranked scores sit on the lane’s scale but never name a winner. Like-for-like rows within 0.5 points read as a practical tie.
Shape of the matched evidence
Only shared public evidence is shown. Sparse evidence stays a ruled list rather than being closed into a radar shape.
A shared-evidence shape is not available.
BenchLM does not draw a radar or infer missing axes when the matched evidence is too sparse.
What each workload costs
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.
Chat turn
1K fresh input + 500 output tokens
- MiniMax M3
- $0.0009
- Fits in one request
- ZAYA1-8B
- Self-hosted; infrastructure cost varies
- Fits in one request
ZAYA1-8B has no comparable published API token rate.
Repository review
50K fresh input + 3K output tokens
- MiniMax M3
- $0.0186
- Fits in one request
- ZAYA1-8B
- Self-hosted; infrastructure cost varies
- Fits in one request
ZAYA1-8B has no comparable published API token rate.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- MiniMax M3
- $0.03
- Fits in one request
- ZAYA1-8B
- Self-hosted; infrastructure cost varies
- Does not fit in one request
- Cached-input rate unavailable
ZAYA1-8B does not fit this workload in one request. ZAYA1-8B has no comparable published API token rate.
Specification differences
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
SpecificationMiniMax M3ZAYA1-8B
Context window
Maximum documented context; output-token limits may be lower.
MiniMax M3
1M
ZAYA1-8B
131K
API model ID
MiniMax M3
Not sourced
ZAYA1-8B
Not sourced
Cached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
MiniMax M3
$0.06 per 1M cached input tokens
ZAYA1-8B
No comparable hosted API rate
Documented inputs
MiniMax M3
Not sourced
ZAYA1-8B
Not sourced
Documented outputs
MiniMax M3
Not sourced
ZAYA1-8B
Not sourced
Provider availability
MiniMax M3
Not sourced
ZAYA1-8B
Not sourced
Reasoning profile
MiniMax M3
Non-Reasoning
ZAYA1-8B
Reasoning
Weight access
MiniMax M3
Open Weight
ZAYA1-8B
Open Weight
License
MiniMax M3
Open Weight
ZAYA1-8B
Open Weight
Release date
MiniMax M3
2026-06-01
ZAYA1-8B
2026-05-05
If you already use one of these models
- Deployment change
- The models list different providers, so authentication, endpoint behavior, limits, and feature support may change.
- Quality signal
- The public evidence has no benchmark result shared by both models, so it does not support a quality verdict.
- Workload cost
- A complete comparable API-rate estimate is not available for both models.
- Context tradeoff
- MiniMax M3 has the larger documented window (1M).
Run the same representative tasks against both endpoints before changing production traffic.
Benchmark evidence
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Browse raw public benchmark evidence38 rows
Agentic
Terminal-Bench 2.0
MiniMax M366%SourceZAYA1-8B—Not directly comparable
BrowseComp
MiniMax M383.5%SourceZAYA1-8B—Not directly comparable
OSWorld-Verified
MiniMax M370.1%SourceZAYA1-8B—Not directly comparable
MCP Atlas
MiniMax M374.2%SourceZAYA1-8B—Not directly comparable
Claw-Eval
MiniMax M374.5%SourceZAYA1-8B—Not directly comparable
BankerToolBench
MiniMax M376.1%SourceZAYA1-8B—Not directly comparable
ResearchClawBench
MiniMax M319.8%SourceZAYA1-8B—Not directly comparable
OSWorld 2.0
MiniMax M34.6%SourceZAYA1-8B—Not directly comparable
Terminal-Bench 2.1 (Vals)
MiniMax M353.6%SourceZAYA1-8B—Not directly comparable
BFCL v4
MiniMax M3—ZAYA1-8B39.2%SourceNot directly comparable
Coding
SWE-bench Verified
MiniMax M380.5%SourceZAYA1-8B—Not directly comparable
SWE-bench Pro
MiniMax M359%SourceZAYA1-8B—Not directly comparable
Terminal-Bench 2.0
MiniMax M366.0%SourceZAYA1-8B—Not directly comparable
NL2Repo
MiniMax M342.1%SourceZAYA1-8B—Not directly comparable
VIBE V2
MiniMax M350.1%SourceZAYA1-8B—Not directly comparable
SVG-Bench
MiniMax M363.7%SourceZAYA1-8B—Not directly comparable
KernelBench Hard
MiniMax M328.8%SourceZAYA1-8B—Not directly comparable
OpenHarmony Bench
MiniMax M348.4%SourceZAYA1-8B—Not directly comparable
LiveCodeBench (Vals)
MiniMax M382.2%SourceZAYA1-8B—Not directly comparable
SWE-bench (Vals)
MiniMax M375.0%SourceZAYA1-8B—Not directly comparable
LiveCodeBench v6
MiniMax M3—ZAYA1-8B65.8%SourceNot directly comparable
Knowledge
GPQA Diamond (Vals)
MiniMax M392.7%SourceZAYA1-8B—Not directly comparable
MMLU-Pro (Vals)
MiniMax M384.2%SourceZAYA1-8B—Not directly comparable
GPQA
MiniMax M3—ZAYA1-8B71%SourceNot directly comparable
GPQA-D
MiniMax M3—ZAYA1-8B71.0%SourceNot directly comparable
MMLU-Pro
MiniMax M3—ZAYA1-8B74.2%SourceNot directly comparable
Math
USAMO 2026
MiniMax M385.7%SourceZAYA1-8B—Not directly comparable
AIME26
MiniMax M3—ZAYA1-8B89.1%SourceNot directly comparable
HMMT Feb 2026
MiniMax M3—ZAYA1-8B71.6%SourceNot directly comparable
IMOAnswerBench
MiniMax M3—ZAYA1-8B59.3%SourceNot directly comparable
Apex
MiniMax M3—ZAYA1-8B32.2%SourceNot directly comparable
Multimodal
OfficeQA Pro
MiniMax M345.1%SourceZAYA1-8B—Not directly comparable
OmniDocBench 1.5
MiniMax M391.6%SourceZAYA1-8B—Not directly comparable
MMMU-Pro
MiniMax M378.1%SourceZAYA1-8B—Not directly comparable
VideoMMMU
MiniMax M384.6%SourceZAYA1-8B—Not directly comparable
Video-MME (with subtitle)
MiniMax M385.4%SourceZAYA1-8B—Not directly comparable
Instruction following
Frequently asked questions
Which is better, MiniMax M3 or ZAYA1-8B?
The public evidence has no benchmark result shared by both models, so it does not support a quality verdict. The page therefore keeps the decision tied to the specific documented workload.
Which is better for coding, MiniMax M3 or ZAYA1-8B?
ZAYA1-8B is not ranked on the public lane for coding, so no winner is named for coding.
Which is better for agentic tasks, MiniMax M3 or ZAYA1-8B?
ZAYA1-8B scores higher for agentic tasks on the public lane, 46.4 to 42. ZAYA1-8B is scored on Estimated evidence for agentic tasks, so the reading is directional rather than like-for-like. BenchLM does not name a winner for agentic tasks from a directional reading; read the shared benchmark rows directly and test the models on the same task set.
Which costs less, MiniMax M3 or ZAYA1-8B?
Both models do not have comparable published API token rates, so this page does not name a universal price winner.
Which has the larger context window, MiniMax M3 or ZAYA1-8B?
MiniMax M3 has the larger documented context window: 1M, compared with 131K.
Related comparisons
Last updated September 10, 2026
Watch MiniMax M3 vs ZAYA1-8B
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.