Skip to content
Token Perks

Menu

Compare the tracked offers

Verified Sep 6–7 2026

Together AI — tracked routes and prices

4 access routes across api (per token), credits / prepaid. Cheapest verified per-token route: Llama 3.3 70B / Llama 3 8B Instruct Lite at $1.04/M blended. Every route was verified directly on its official page this pass. Snapshot 2026-09-07 (rows accessed 2026-09-07). Rankings live on the cost leaderboard.

  • Qwen3.8-2.4T-A95B / Qwen3.8 Flash / Qwen3.7-Max / Qwen3.7-Plus

    $2.00/$6.00; $0.15/$0.47; $1.25/$3.75; $0.32/$1.28

    API (per token) · DIRECT · Resale pricing, not first-party

  • Qwen3.6-Plus / Qwen3.5-397B-A17B / Qwen3.5-9B / Qwen3-235B-A22B

    $0.50/$3.00; $0.60/$3.60; $0.17/$0.25; $0.20/$0.60

    API (per token) · DIRECT · Resale pricing, not first-party

  • Llama 3.3 70B / Llama 3 8B Instruct Lite

    $1.04/$1.04; $0.14/$0.14

    API (per token) · DIRECT · Llama 4 pricing unverified

  • Serverless PAYG (no credit packs)

    Per-token postpaid; H100 dedicated $3.99/hr; B200 $8.19/hr

    Credits / prepaid · DIRECT

Route economics at a glance

Assembled from the verified rows above — no new data. Each figure links its official source and carries the date it was read.

Together AI economics summary computed from tracked rows
Free / promo routes0None sold by Together AI directly.
API per-token range (blended)$1.04–$3.00/MAcross 3 priced API routes; blended = (3 x input + output) / 4 — see methodology.
Batch / off-peak discounts0None published on tracked routes.
Published overage terms0 of 1 routesChecked — no per-unit figure published on tracked routes.

Cited intelligence scores

Quoted from the Artificial Analysis Intelligence Index v4.3, accessed 2026-09-07 — not measured by us. Each value links its exact AA source row. Artificial Analysis scores every reasoning-effort variant separately; we quote the named variant shown. An asterisk marks scores AA itself flags as estimates.

  • Qwen3.8 2.4T A95B 40

    AA Intelligence Index v4.3 — Source: Artificial Analysis, accessed 2026-09-07 · AA source · priced on this page as Qwen3.8-2.4T-A95B / Qwen3.8 Flash / Qwen3.7-Max / Qwen3.7-Plus

  • Llama 4 Maverick 9

    AA Intelligence Index v4.3 — Source: Artificial Analysis, accessed 2026-09-07 · AA source · priced on this page as Llama 3.3 70B / Llama 3 8B Instruct Lite

Sources & methods

Snapshot 2026-09-07 (per-datum access dates at right; a few rows rest on dated official snapshots — see notes). Every price above links its official source in the routes table; every score above links its AA source row. Below, each datum's source and access date, in full. Method: methodology.

Together AI datum-level sources: each price and score with its source and access date
DatumSourceAccessedEvidence
Qwen3.8-2.4T-A95B / Qwen3.8 Flash / Qwen3.7-Max / Qwen3.7-PlusAPI (per token) · $2.00/$6.00; $0.15/$0.47; $1.25/$3.75; $0.32/$1.28together.ai2026-09-07DIRECT
Qwen3.6-Plus / Qwen3.5-397B-A17B / Qwen3.5-9B / Qwen3-235B-A22BAPI (per token) · $0.50/$3.00; $0.60/$3.60; $0.17/$0.25; $0.20/$0.60together.ai2026-09-07DIRECT
Llama 3.3 70B / Llama 3 8B Instruct LiteAPI (per token) · $1.04/$1.04; $0.14/$0.14together.ai2026-09-07DIRECT
Serverless PAYG (no credit packs)Credits / prepaid · Per-token postpaid; H100 dedicated $3.99/hr; B200 $8.19/hrtogether.ai2026-09-07DIRECT
Qwen3.8 2.4T A95B — II v4.3: 40Quoted score, not measured by usartificialanalysis.ai2026-09-07CITED
Llama 4 Maverick — II v4.3: 9Quoted score, not measured by usartificialanalysis.ai2026-09-07CITED