ModelPriceWatch.com
Last scan 2026-09-11 Models tracked 260 Providers 32 Cheapest paid Granite 4.0 H Micro $0.017/Mtok in Every price links to its source

KAT-Coder-Air V2.5 vs Mercury Edit 2

Side-by-side comparison of API pricing, specs, benchmarks, and capabilities

KAT-Coder-Air V2.5 is 30.9% cheaper on blended cost ($0.259 vs $0.375/Mtok)
Specification
Overview
StatusCurrent coding Current coding
Released
Pricing per million tokens
Input $0.148/Mtok $0.250/Mtok
Output $0.593/Mtok $0.750/Mtok
Blended avg $0.259/Mtok $0.375/Mtok
Cached input $0.030/Mtok $0.025/Mtok
Price basis USD is FX-derived from StreamLake's RMB list price (¥1 in / ¥0.2 cached / ¥4 out per 1M) at 6.7476 CNY/USD (2026-08-07); re-checked on FX drift, not a promo. Inception's own rate card at docs.inceptionlabs.ai/get-started/models prices Mercury Edit 2 at $0.25 input / $0.025 cached input / $0.75 output per 1M tokens, in a table headed "Input Price (1M Tokens)", so the per-1M basis is the one the vendor declares. Unlike the Mercury 2.5 row on the same card, this row carries no strikethrough and no discount marker: the 80%-off banner on that page names Mercury 2.5 only, so these are plain list prices and not a promotional rate. Confirmed on a second first-party surface the same day: the API's edit- and FIM-scoped model lists (api.inceptionlabs.ai/v1/edit/completions/models and /v1/fim/completions/models) both serve mercury-edit-2 at prompt 0.00000025 and completion 0.00000075 per TOKEN with input_cache_reads 0.000000025 and cache writes free, which is the same $0.25 / $0.025 / $0.75 per 1M. The chat-scoped list at api.inceptionlabs.ai/v1/models does NOT carry this SKU — it lists Mercury 2 and Mercury 2.5 only — so that route's silence is about the endpoint it serves, not about the model. Verified 2026-09-11. Context window, on those same two surfaces, is the one field where they disagree, and we publish the rate card's figure: the card prints "FIM: 32K / NextEdit: 32K" with 8,192 max output tokens, stated per endpoint, while the model-list JSON reports context_length 128000 and max_output_length 32000. That JSON record reads as a template copied from Mercury 2 — 128000 is Mercury 2's exact window, its description is Mercury 2's text, and its created timestamp is byte-identical to the one served for Mercury 2.5, a model Inception launched on 2026-09-08 — so the card, which states a window per endpoint, is the more specific claim. None of the three prices is affected: both surfaces agree on all of them to the digit.
Specifications
Context window 256K tokens 32K tokens
Parameters Proprietary Proprietary
Speed (TPS)
Modalities
Input
text
text
Providers
Available from
Kwaipilot — $0.148/$0.593/Mtok
Inception — $0.250/$0.750/Mtok

Cost at scale

1M tokens · 50/50 input/output
Projected cost of KAT-Coder-Air V2.5 vs Mercury Edit 2 at increasing token volumes
VolumeKAT-Coder-Air V2.5Mercury Edit 2Savings
1M tokens $0.26 $0.38 $0.12 (32%)
10M tokens $2.59 $3.75 $1.16 (30.9%)
100M tokens $25.92 $37.5 $11.58 (30.9%)
1000M tokens $259.25 $375 $115.75 (30.9%)

When to pick which

distilled from the pricing and spec data above
Pick KAT-Coder-Air V2.5 if…
  • cost dominates: $0.259/Mtok blended vs $0.375 — 30.9% less on the same 50/50 token mix
  • you need the longer context: 256K tokens vs 32K (8×)
Pick Mercury Edit 2 if…
  • your workload re-reads context (agents, RAG, long chats): cached input costs $0.025/Mtok — 90% off its list input price

Summary

KAT-Coder-Air V2.5 by Kwaipilot costs $0.148/Mtok input and $0.593/Mtok output, with a 256K-token context window. It supports text input.

Mercury Edit 2 by Inception costs $0.250/Mtok input and $0.750/Mtok output, with a 32K-token context window. It supports text input.

On a blended cost basis, KAT-Coder-Air V2.5 is 30.9% cheaper than Mercury Edit 2.It also has a larger context window.

Note: Pricing is per million tokens. Actual costs vary with usage patterns, prompt caching, and batch discounts. Always verify against official provider pricing pages.

More comparisons

pairs sharing a model with this page