Skip to content
Heimdallm
Marketplace/Model library
XiaomiModel family2 served versions

Xiaomi: MiMo V2.6 Flash RL Abliterated (DealignAI)

xiaomimimo/mimo-v2.6-flash-rlHugging FaceChatVisionmoereasoningtoolsabliterated

Input
$0.008 – $0.014
Per million input tokens
Output
$0.017 – $0.028
Per million output tokens
Context
1M
Largest endpoint
Max output
66K
Largest endpoint
Modality
Text + Image → Text
textimage

Versions

Every served version of this family. Each has its own providers and prices.

Providers

Live endpoints for this model, cheapest first. TTFT is time to first token, measured on streamed requests; response is request to last byte, measured on every request; throughput covers the last day; uptime is measured over the last day and week.

ProviderContextMax outInputOutputCache readTTFTResponseThroughputCapacityUptime 1dUptime 7dToolsReasoningHealthData
Consensus Protocolfp890% offbaseUS131K33K$0.008$0.017$0.002654ms30m3.40s30m61 tok/s8/8 free99.5%99.3%VerifiedEffortHealthyZDR
ImpCCmxfp490% offmimo-v2.6-flash-rl-abliteratedUS1M66K$0.014$0.028$0.00028341ms30m4.94s30m88 tok/s8/8 free100.0%95.8%VerifiedEffortHealthyZDR

Pricing

USD per million tokens.

ProviderInputOutputCache read
Consensus Protocolfp890% offbase$0.008$0.017$0.002
ImpCCmxfp490% offmimo-v2.6-flash-rl-abliterated$0.014$0.028$0.00028

Price history

Input price · USD per million tokens

Output price · USD per million tokens

Performance

Timing percentiles cover the labelled window: the last 30 minutes when it has enough samples, otherwise the last day. TTFT is the seconds to the first token, measured only on streamed requests; response is the seconds from request to last byte, measured on every request, streamed or not. Throughput is output tokens per second over the last day, counted only over streamed generations long enough to have a rate — a buffered answer has no generation phase, and a two-token one times the scheduler rather than the model. Tool calls is the share of the last day of tool requests this endpoint answered with a tool call, published once the day holds twenty of them; under that the cell is blank rather than a rate nothing measured.

ProviderTTFT p50p75p90p99Response p50p75p90p99Throughput p50p75p90p99Tool calls 1dRunaways 1d
Consensus Protocolfp8base654ms30m1.02s3.49s5.16s3.40s30m6.18s35.9s164.3s61 tok/s69 tok/s71 tok/s72 tok/s—0.8%
ImpCCmxfp4mimo-v2.6-flash-rl-abliterated341ms30m612ms1.22s1.24s4.94s30m7.21s24.3s121.3s88 tok/s99 tok/s109 tok/s114 tok/s—0.0%

Time to first token by hour · last 7 days

Hourly median per provider. The line spans hours with no streamed traffic; where samples are sparse, a dot marks each hour actually timed.

Throughput by hour · last 7 days

Hourly median output tokens per second per provider, over generations long enough to have a rate. The line spans hours with no answer long enough to measure; where samples are sparse, a dot marks each hour actually measured.

Uptime

Share of successful requests and probes over the last day and week.

Availability by hour · last 7 days

Share of successful requests and probes per hour and provider. The line spans hours with nothing to measure; where samples are sparse, a dot marks each hour actually measured.

Activity

Prompt tokens and completion tokens per UTC day over the last 30 days. Every billed request counts, the platform's own paid probe traffic included.

Prompt tokens36.7Mlast 30 days
Completion tokens11.9Mlast 30 days
Quickstart
curl https://api-staging.heimdallm.com/api/v1/chat/completions \  -H "Authorization: Bearer $HEIMDALL_API_KEY" \  -H "Content-Type: application/json" \  -d '{    "model": "xiaomimimo/mimo-v2.6-flash-rl",    "messages": [{ "role": "user", "content": "Hello" }]  }'

Set HEIMDALL_API_KEY to a key from your account. See the quickstart for routing suffixes and provider preferences.

About

Author
Xiaomi
Family id
xiaomimimo/mimo-v2.6-flash-rl

XiaomiMiMo: MiMo-V2.6-Flash-RLReleased Sep 21, 2026

Scaling Reinforcement Learning Toward Self-Improvement

Parameters
frequency_penalty (-2–2), max_tokens (1–32768), min_p (0–1), presence_penalty (-2–2), reasoning_effort (none, low, medium, high), repetition_penalty (0–2), seed (0–2147483647), stop (list), temperature (0–2), tool_choice (any value), tools (true/false), top_k (0–200), top_p (0–1)
Licence
mit
Size
310.8B parameters, U8 weights
Tokenizer
—
Available until
—

Xiaomi: MiMo V2.6 Flash RL Abliterated (DealignAI)Released Sep 22, 2026

DealignAI's abliterated MiMo V2.6 Flash RL supports text and image input, reasoning, tool calling and structured JSON output. Each request has a 1,000,000-token context limit with at most 65,536 output tokens.

Parameters
chat_template_kwargs (object), frequency_penalty (-2–2), guided_choice (list), guided_grammar (text), guided_json (any value), guided_regex (text), ignore_eos (true/false), logprobs (true/false), max_completion_tokens (1–65536), max_tokens (1–65536), min_tokens (0–65536), n (1–4), parallel_tool_calls (true/false), presence_penalty (-2–2), reasoning (object), reasoning_effort (none, medium), repetition_penalty (>0, ≤2), response_format (object), seed (integer), stop (text or list), structured_outputs (true/false), temperature (0–2), tool_call_dispatch (true/false), tool_choice (none, auto), tools (list), top_k (-1–2147483647), top_logprobs (0–20; requires logprobs=true), top_p (>0, ≤1)
Licence
mit
Size
159.4B parameters, U8 weights
Tokenizer
—
Available until
—

From the model card

What the repository states about these weights, read from XiaomiMiMo/MiMo-V2.6-Flash-RL on Oct 8, 2026.

Languages
English, Chinese
Papers
—
Datasets
—
Revision
5711b26Sep 22, 2026
Downloads
54.6K
Likes
544
DiscoverFind your next modelModel libraryAll models, versions, prices and capabilitiesBenchmarksMeasured quality, cost and time by serving configurationShowcaseImages, video and 3D made on the marketplaceProvidersCompare independent inference providersActivityMarketplace traffic and usageNetwork statusAvailability, latency and endpoint healthChatConversations, model settings and filesImage, video & 3DGenerate images, video and 3D assetsScene studioStage a shot in 3D, direct it and render it as a videoCoding playgroundBuild software with an AI agent on a machine of your ownOverview & routingIdentity, workspaces and routing preferencesCredits & billingBalance, deposits, ledger and reserved fundsAPI keysKeys scoped to workspaces, limits and expirationOrganizationsTeams, members and shared workspacesRequest historyRequests, receipts, sandbox sessions and account historyReferralsInvitations and referral earningsPrivacyData settings, exports and deletionDocumentationQuickstart, SDKs, pricing and guidesAPI referenceInteractive endpoint reference