Data as of November 5, 2026 · How the score is built
Released Sep 22, 20261M contextXiaomi MiMo-V2.6-Pro-RL model card
MiMo-V2.6-Pro
Decision readingMiMo-V2.6-Pro scores 57.3 out of 100 and ranks #58 of 216. This profile shows 16 source-displayable benchmark rows; its strongest eligible category is Coding at #25. API pricing is $0.435 input and $0.87 output per million tokens, with cached input at $0.0036.
Released Sep 22, 2026 — see all recent releases
Decision snapshot
Each value carries a field reference instead of floating alone. Markers compare this model with the current ranked and priced catalog; they are not absolute quality thresholds.
Capability
57.3/100
field median 50.1#58 of 216 ranked models
Public
#58of 216
Verified —
Price
$0.43input / $0.87 output
input median $0.65cached $0.004 · blended $0.65
Speed
66tok/s
field median 92 tok/sFirst token 33.07 s
Context
1Mtokens
field median 256,000Maximum output length is tracked separately
Strongest published evidence
Coding ranks #25. Particularly well-suited for software development and code generation tasks.
Validate before choosing
16 published rows leave some tracked benchmark slots empty. Knowledge is its lowest eligible category at #42.
Source-linked · 16 displayable benchmark rows
Category score record
Scores and ranks appear only where published evidence can be displayed. The table keeps the score, weight, cohort, and evidence state together.
| Category | Score | Rank | Percentile | Weight | Benchmarks | Evidence |
|---|---|---|---|---|---|---|
| AgenticRank #40 of 135Percentile 71stWeight 22%9 benchmarksVerified | 56.6 | 9 benchmarks | Verified | |||
| CodingRank #25 of 156Percentile 85thWeight 20%4 benchmarksVerified | 57.1 | 4 benchmarks | Verified | |||
| ReasoningRank Not rankedWeight 17%1 benchmarkVerified | 82.5 | 1 benchmark | Verified | |||
| MultimodalWeight 12%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| KnowledgeRank #42 of 177Percentile 77thWeight 12%Not listedUnlisted sources | 62.3 | Not listed | Unlisted sources | |||
| MultilingualWeight 7%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| Inst. FollowingWeight 5%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| MathWeight 5%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured |
16 of 802 tracked benchmark slots have displayable evidence · bars run 0–100
Coverage detailsHow much of this is verified
Coverage is split by category so a strong number never hides a thin evidence base. Verified means the row is tied to a published source; provisional rows remain visible but separate.
- Agentic9/9 verified
- Coding4/4 verified
- Reasoning1/1 verified
- MultimodalNot measured
- KnowledgeNot measured
- MultilingualNot measured
- Inst. FollowingNot measured
- MathNot measured
Capability shape
Each axis shows percentile within that category’s eligible cohort. The comparison outline is the median of the six nearest public-score peers; a collapsed vertex means the category is not rank-eligible.
MiMo-V2.6-Pro category percentile values
- Agentic71st percentile
- Coding85th percentile
- ReasoningNot eligible
- MultimodalNot eligible
- Knowledge77th percentile
- MultilingualNot eligible
- Instruction followingNot eligible
- MathNot eligible
The dashed outline is median of 6 nearest peers.
Eligible category ranks
- Agentic#40/135
- Coding#25/156
- ReasoningNot ranked
- MultimodalNot ranked
- Knowledge#42/177
- MultilingualNot ranked
- Inst. FollowingNot ranked
- MathNot ranked
Benchmark ledger
Coding opens by default. The marker compares each value with the best source-verified result in the catalog; provisional leaders do not set the reference. Expand the remaining categories for every published row.
Coding4 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| DeepSWE | Score71.9% | Versus best verified row Best verified: Muse Wonder Full · 79.0% | Gap7.1 behind | Weight15% ref. weight | Provider exact |
| Terminal-Bench 2.1 | Score89.9% | Versus best verified row Best verified: SWE-2 · 92.8% | Gap2.9 behind | WeightScored in Agentic | Provider exact |
| ProgramBenchProgramBench: Can Language Models Rebuild Programs From Scratch? | Score26.5% | Versus best verified row Best verified: Claude Opus 5 · 93.0% | Gap66.5 behind | WeightDisplay only | Provider exact |
| Bug Hunt Bench | Score22.7 fixes | Versus best verified row Best verified: Claude Sonnet 5.5 · 51.3 fixes | Gap28.6 behind | WeightDisplay only | Benchmark exact |
Agentic9 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| Terminal-Bench 4.0 | Score34.90% | Versus best verified row Best verified: Claude Sonnet 5.5 · 70.60% | Gap35.7 behind | Weight8% ref. weight | Provider exact |
| Terminal-Bench 2.1 | Score89.9% | Versus best verified row Best verified: SWE-2 · 92.8% | Gap2.9 behind | Weight8% ref. weight | Provider exact |
| OSWorld-Verified | Score82% | Versus best verified row Best verified: Qwen3.8 Max · 86.1% | Gap4.1 behind | Weight7% ref. weight | Provider exact |
| AutomationBench | Score53.1% | Versus best verified row Best verified: MiniMax M3.1 Flash · 62.2% | Gap9.1 behind | Weight6% ref. weight | Provider exact |
| JobBench | Score62.0% | Versus best verified row Best verified: Muse Wonder Full · 75.3% | Gap13.3 behind | Weight6% ref. weight | Provider exact |
| Toolathlon-Verified | Score76.9% | Versus best verified row Best verified: Claude Opus 5 · 80.6% | Gap3.7 behind | Weight3% ref. weight | Provider exact |
| Agents' Last Exam | Score31.6% | Versus best verified row Best verified: GPT-6.1 Terra · 60.0% | Gap28.4 behind | Weight3% ref. weight | Provider exact |
| CyberGym | Score94.0% | Versus best verified row Best verified: MiMo-V2.6-Flash · 95.1% | Gap1.1 behind | WeightDisplay only | Provider exact |
| ExploitGym | Score17.8% | Versus best verified row Best verified: Gemini 4 Carbon · 53.5% | Gap35.7 behind | WeightDisplay only | Provider exact |
Reasoning1 row
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| STEMBenchSTEMBench120: six-domain scientific research progress | Score2.17 | Versus best verified row Best verified: GPT-6 Astra · 2.81 | Gap0.6 behind | WeightDisplay only | Benchmark exact |
External signals2 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| ExploitBenchExploitBench v8-bench | Score48% | Versus best verified row Best verified: GPT-6 Astra · 100% | Gap52.1 behind | WeightDisplay only | Provider exact |
| SEC-Bench Pro | Score66.3% | Versus best verified row Best verified: Gemini 4 Carbon · 88.2% | Gap21.9 behind | WeightDisplay only | Provider exact |
What it costs to get this score
Published API price against the public score. The x-axis uses a log scale; the dashed path marks models that are not beaten by a cheaper, higher-scoring option. Price uses average of published input and output rates.
MiMo-V2.6-Pro · 57.3 score · $0.65 blended per million tokens
The chart opens on the current model. Scroll horizontally to inspect the full price axis.
Horizontal: blended price per million tokens, log scale · Vertical: public score
Lineage
The sequence follows explicit supersedes links. Each score is estimated for that model; a relative can inform a sparse estimate but never sets a floor, so a newer release can score below an earlier one. Scores and prices remain blank when the corresponding public row or first-party rate is unavailable.
Spec sheet
Each documented value carries its source. Missing fields stay visible as not sourced or not published, rather than disappearing from the page.
- API model ID
- Not publishedXiaomi MiMo-V2.6 launch
- Context window
- 1MXiaomi MiMo-V2.6 launch
- Maximum output
- Not sourced yet
- Knowledge cutoff
- Not sourced yet
- Input modalities
- Not sourced yet
- Output modalities
- Not sourced yet
- Parameters
- Not sourced yet
- Availability
- Open-weight checkpoint and hosted API are available at launch. Xiaomi lists text, image, video, and audio input modalities.
- Cloud regions
- Not tracked yet
- Lifecycle
- Current
- API capabilities
- Tool calling, structured outputs, and batch support are not tracked yet
- Prompt caching
- Published at $0.004 per million cached input tokensXiaomi MiMo-V2.6 launch
- Self-host
- MiMo-V2.6-Pro needs ~1128GB VRAM (8× NVIDIA H200 (141GB)).
- Rate limits
- Not tracked yet
How to read this profile
The visual layer above carries the decisions. These notes preserve the model, ranking, coverage, and family context behind the numbers.
MiMo-V2.6-Pro ranks #58 of 216 on the public leaderboard with a score of 57.27/100. It does not yet have enough sourced coverage for a verified position.
MiMo-V2.6-Pro is a open weight model with a 1M context window. It uses an explicit reasoning mode, which can improve complex problem solving while adding latency and token use.
Open-weight checkpoint and hosted API are available at launch. Xiaomi lists text, image, video, and audio input modalities.
Official exact-value snapshot from Xiaomi's September 22, 2026 MiMo-V2.6 launch page and technical report. Xiaomi releases the MIT-licensed MiMo-V2.6-Pro-RL checkpoint as a natively omnimodal model. BenchLM maps only directly comparable public benchmarks and keeps the provider-run row non-ranked because weighted benchmark coverage remains sparse. MiMo Code Bench, MiMo Visual Coding, and MiMo Cyber Bench are in-house evaluations and remain outside the shared benchmark schema.
MiMo-V2.6-Pro sits in the MiMo-V2.6 family with MiMo-V2.6-Flash. Its explicit predecessor is MiMo-V2.5-Pro. 16 of 802 tracked benchmark slots currently have displayable evidence. Missing categories stay blank.
Its strongest eligible category is Coding at #25, while its lowest eligible position is Knowledge at #42. particularly well-suited for software development and code generation tasks.
Last updated November 5, 2026. Runtime fields remain blank until a sourced snapshot exists.
Deployment options
Self-host and provider-specific paths stay separate from benchmark evidence so operating constraints are visible before a score becomes the whole decision.
Where to run MiMo-V2.6-Pro by API
3 hosts publish a per-token price for it. DeepInfra has the lowest output rate, $0.78 per million tokens. Prices read from each host’s own pages on October 20, 2026.
| Host | Input /1M | Output /1M | Cached /1M | Context |
|---|---|---|---|---|
| DeepInfra | $0.39 | $0.78 | $0.003 | 1.05M |
| Novita AI | $0.43 | $0.87 | $0.004 | 1.05M |
| OpenRouter | $0.43 | $0.87 | $0.004 | 1.05M |
Self-host vs API cost
Estimates at 50,000 req/day · 1000 tokens/req average.
Questions
How does MiMo-V2.6-Pro perform overall in AI benchmarks?
MiMo-V2.6-Pro ranks #58 out of 216 models on the public BenchAlign leaderboard, with a score of 57.27/100. Its evidence status is Estimated, and this profile shows 16 source-displayable benchmark rows. The label describes evidence depth, not a provider quality claim; inspect category rows before choosing a workload.
Is MiMo-V2.6-Pro good for coding and programming?
MiMo-V2.6-Pro ranks #25 out of 156 eligible models for coding and programming, with a public category score of 57.1/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is MiMo-V2.6-Pro good for reasoning and logic?
MiMo-V2.6-Pro has source-displayable benchmark coverage for reasoning and logic, but the public category table does not assign it a rank there. The individual rows remain available for inspection. A missing category position means the evidence threshold was not met; it does not convert the model's unmeasured work into a zero.
Is MiMo-V2.6-Pro good for agentic tool use and computer tasks?
MiMo-V2.6-Pro ranks #40 out of 135 eligible models for agentic tool use and computer tasks, with a public category score of 56.6/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is MiMo-V2.6-Pro open source?
MiMo-V2.6-Pro is an open-weight model from Xiaomi. Its weights can be downloaded for local or hosted deployment, subject to the published license. Open weight does not automatically mean open source: training data and training code may remain private, and commercial restrictions can still apply.
Which sibling models are related to MiMo-V2.6-Pro?
MiMo-V2.6-Pro belongs to the MiMo-V2.6 family. Related tracked variants include MiMo-V2.6-Flash. A sibling link indicates shared lineage or a documented configuration relationship; it does not mean the variants have identical pricing, context limits, benchmark evidence, or deployment behavior. Compare before switching.
Does MiMo-V2.6-Pro have full benchmark coverage on BenchLM?
No. MiMo-V2.6-Pro currently has 37 source-displayable rows. The profile exposes published, non-generated evidence and leaves missing categories blank until an exact evaluation is available. Coverage describes how much was measured; it is not a penalty added to an individual benchmark result.
What is the context window size of MiMo-V2.6-Pro?
MiMo-V2.6-Pro has a documented context window of 1M. That figure is the maximum combined prompt and retained-conversation space reported for this exact model; it is not the maximum output length. The profile keeps output limits separate because providers often publish those limits independently.