Fugu Ultra v2.0 vs MiMo-V2.5

Sakana AI · Global  |  Xiaomi · China · Updated June 2026

Quick verdict

Pick Fugu Ultra v2.0 for sakana's own evaluation reports best-or-joint-best results on 5 of 8 benchmarks tested, including a 74.3 on the deepswe coding-repair benchmark or claims to outperform flagship models like gpt-6 astra and claude fable 5.1 despite its underlying orchestration pool reportedly excluding both of them. Pick MiMo-V2.5 for native omnimodal — strong image and video understanding or very low cost (~half the inference of the pro tier). Choose MiMo-V2.5 if you need self-hosting or data privacy; Fugu Ultra v2.0 if you want a managed API.

Fugu Ultra v2.0 (Sakana AI) and MiMo-V2.5 (Xiaomi) are two of the models people most often weigh against each other in 2026. Fugu Ultra v2.0 is sakana AI's flagship multi-agent orchestration model, claiming best-or-joint-best results on 5 of 8 self-reported benchmarks against monolithic frontier models — $5/$30 per million tokens. MiMo-V2.5 is xiaomi's cheap omnimodal model — Pro-level agentic perception across image and video at a fraction of the cost. They diverge most on price and open vs. closed weights — each quantified below from the models' real specs.

Key differences at a glance

Side-by-side specs

SpecFugu Ultra v2.0MiMo-V2.5
ProviderSakana AI (Global) Xiaomi (China)
ReleasedSeptember 10, 2026 April 22, 2026
Context window1M tokens (~1,500 pages) 1M (~1,500 pages)
Price (in/out)$5/$30 per 1M tokens $0.14/$0.28 per 1M tokens
Open weight?No — API only Yes — self-hostable
Modalitiestext text, image, audio, video, code
SWE-Bench VerifiedNot published Not published
MRCR v2 @ 1MNot published Not published

Who wins what

Sakana's own evaluation reports best-or-joint-best results on 5 of 8 benchmarks tested, including a 74.3 on the DeepSWE coding-repair benchmark

Fugu Ultra v2.0

Sakana AI's flagship multi-agent orchestration model, claiming best-or-joint-best results on 5 of 8 self-reported benchmarks against monolithic frontier models — $5/$30 per million tokens — and it is the newer of the two.

Claims to outperform flagship models like GPT-6 Astra and Claude Fable 5.1 despite its underlying orchestration pool reportedly excluding both of them

Fugu Ultra v2.0

Fugu Ultra v2.0 lists claims to outperform flagship models like GPT-6 Astra and Claude Fable 5.1 despite its underlying orchestration pool reportedly excluding both of them among its strengths; MiMo-V2.5 does not.

1M-token context window

Fugu Ultra v2.0

Fugu Ultra v2.0 lists 1M-token context window among its strengths; MiMo-V2.5 does not.

Native omnimodal — strong image and video understanding

MiMo-V2.5

Xiaomi's cheap omnimodal model — Pro-level agentic perception across image and video at a fraction of the cost — and it runs cheaper at $0.14/$0.28 per 1M tokens.

Very low cost (~half the inference of the Pro tier)

MiMo-V2.5

At $0.14/$0.28 per 1M tokens it undercuts Fugu Ultra v2.0 ($5/$30 per 1M tokens), and that gap compounds at volume.

Agent-framework integration

MiMo-V2.5

Fugu Ultra v2.0 is comparatively weak here — not a single trained model — Sakana describes Fugu as "a Multi-Agent System, Delivered as One Model" that dynamically routes tasks across a pool of other models, so head-to-head benchmark framing against monolithic frontier models should be read skeptically

Lowest cost at scale

MiMo-V2.5

At $0.14/$0.28 per 1M tokens, it is the cheaper of the two — the gap dominates the bill on high-volume workloads.

Which should you pick?

A cost-sensitive startup shipping high volume

MiMo-V2.5

At $0.14/$0.28 per 1M tokens it undercuts Fugu Ultra v2.0, and on millions of tokens that margin decides the monthly bill.

A team with data-privacy or self-hosting needs

MiMo-V2.5

Open weights let you run it on your own hardware; Fugu Ultra v2.0 is API-only.

Anyone whose priority is sakana's own evaluation reports best-or-joint-best results on 5 of 8 benchmarks tested, including a 74.3 on the deepswe coding-repair benchmark

Fugu Ultra v2.0

It is specifically built for that.

Anyone whose priority is native omnimodal — strong image and video understanding

MiMo-V2.5

That is its strongest area.

Fugu Ultra v2.0: where it fits

Sakana AI's flagship multi-agent orchestration model, claiming best-or-joint-best results on 5 of 8 self-reported benchmarks against monolithic frontier models — $5/$30 per million tokens. Released September 10, 2026 by Sakana AI, it is built for sakana's own evaluation reports best-or-joint-best results on 5 of 8 benchmarks tested, including a 74.3 on the DeepSWE coding-repair benchmark, claims to outperform flagship models like GPT-6 Astra and Claude Fable 5.1 despite its underlying orchestration pool reportedly excluding both of them, and 1M-token context window.

Its trade-offs are real: not a single trained model — Sakana describes Fugu as "a Multi-Agent System, Delivered as One Model" that dynamically routes tasks across a pool of other models, so head-to-head benchmark framing against monolithic frontier models should be read skeptically, pricing rises to roughly $10/$45 per million tokens for prompts above 272K tokens, and benchmark comparisons against GPT-6 Astra and Claude Fable 5.1 are Sakana's own self-reported figures, not independently verified head-to-head scores. At $5 in / $30 out per million tokens, it sits in the premium price band.

MiMo-V2.5: where it fits

Xiaomi's cheap omnimodal model — Pro-level agentic perception across image and video at a fraction of the cost. Released April 22, 2026 by Xiaomi, it is built for native omnimodal — strong image and video understanding, very low cost (~half the inference of the Pro tier), agent-framework integration, and 1M context for full documents in one pass.

Its trade-offs: not the deepest reasoning tier (see V2.5-Pro), and limited Western tooling and support. At $0.14 in / $0.28 out per million tokens, it sits in the budget price band.

The bottom line for this matchup

The defining split here is open vs. closed. MiMo-V2.5 gives you weights you control — self-host it, fine-tune it, keep data in-house, pay only for hardware. Fugu Ultra v2.0 gives you a managed, always-updated API with no infrastructure to run. Teams with GPUs, privacy requirements, or huge volume often favour the open model; teams that want zero ops and the latest capabilities favour the closed one. Capability is close enough that this operational question, not the benchmark, usually decides it.

Want both Fugu Ultra v2.0 and MiMo-V2.5 without two subscriptions? LumiChats gives you these plus 40+ models under one ₹69/day pass (about $1/day) — draft with one, cross-check with the other.

See pricing

Frequently asked questions

Is Fugu Ultra v2.0 or MiMo-V2.5 better for coding?

Public SWE-Bench figures are not available for either model, so the honest test is your own repository — run an identical real bug through both. By design, Fugu Ultra v2.0 leans toward sakana's own evaluation reports best-or-joint-best results on 5 of 8 benchmarks tested, including a 74.3 on the deepswe coding-repair benchmark while MiMo-V2.5 leans toward native omnimodal — strong image and video understanding, and that positioning usually predicts which feels better on your codebase.

Which is cheaper, Fugu Ultra v2.0 or MiMo-V2.5?

MiMo-V2.5 is open-weight, so self-hosting means no per-token fee (you pay for hardware instead), while Fugu Ultra v2.0 is API-metered at $5/$30 per 1M tokens. For most teams without GPUs, the API model is cheaper to start; at very high volume, self-hosting can win.

Which has the bigger context window?

Both advertise 1M tokens (~1,500 pages). Remember advertised ≠ usable: recall typically degrades before the ceiling.

Can I use both Fugu Ultra v2.0 and MiMo-V2.5 together?

Yes — a multi-model platform like LumiChats gives you Fugu Ultra v2.0, MiMo-V2.5 and 40+ others under one ₹69/day pass (about $1/day), so you can draft with one and cross-check with the other instead of buying two subscriptions.

Which is newer, Fugu Ultra v2.0 or MiMo-V2.5?

Fugu Ultra v2.0 — released September 10, 2026, about 5 months after MiMo-V2.5.

Related comparisons

Specifications and benchmarks reflect publicly reported figures as of June 2026 and may change as providers release updates. Always verify on your own workload.